JPH11317961A - Image coder, image decoder, image coding method, image decoding method and data storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reference color motion vector corresponding to an interlace color signal even in prediction processing for a shape motion vector in a shape coder where motion compensation coding processing is conducted at all times in unit of frames. SOLUTION: The image coder that selects adaptively processing in unit of frames or processing in unit of fields as motion compensation coding processing with respect to an interlace image signal corresponding to an object is provided with an MV prediction device 204a that generates a predicted value of a shape motion vector by referencing a color motion vector. In the case that the color motion vector to be referenced corresponds to the motion coding processing in unit of fields, the MV prediction device 204a references an output of a frame MV converter 110b that converts a color motion vector in unit of fields into a motion vector in unit of frames.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image encoding apparatus and an image decoding apparatus, an image encoding method and an image decoding method, and a data storage medium, and more particularly to a shape of an object constituting an image on one screen. Apparatus and method for efficiently encoding a motion vector related to a shape signal indicating shape, apparatus and method for decoding an encoded signal obtained by encoding the shape motion vector, and encoding processing or shape of the shape motion vector The present invention relates to a data storage medium storing an image processing program for performing a motion vector decoding process by software.

[0002]

2. Description of the Related Art In recent years, a multimedia era has been entered in which voice, image, and other data are handled in an integrated manner. Has become a feature of multimedia. Generally, multimedia means not only characters but also graphics, sounds, and especially images, etc. are simultaneously associated with each other. To make the above-mentioned conventional information media the target of multimedia, the information must be expressed in digital form. Is an essential condition.

However, when the amount of information handled by each of the above information media is estimated as a digital information amount, the amount of information per character is 1 to 2 bytes in the case of characters, whereas the amount of information per character is 1 to 2 bytes in the case of voice. 64Kbits (telephone quality), plus 100Mbits per second for video
(Current television broadcast quality) or more is required, and in most of the above-described information media, it is not realistic to handle the vast amount of information in a digital form as it is. For example, a videophone is 64Kbps to 1.5Mbp.
service integrated digital network (ISD)
N: Integrated Services Digital Network), but it is impossible to send video information of a television camera as it is via ISDN.

Therefore, what is needed is an information compression technique. For example, in the case of a videophone, ITU-T
(International Telecommunication Union Telecommunication Standardization Sector) 261 and H.E. H.263 video compression technology is used. According to the information compression technology of the MPEG1 standard, a normal music CD (compact disc) is used.
It is also possible to include image information together with audio information.

Here, MPEG (Moving Picture Exper)
ts Group) is an international standard relating to a technique for compressing moving image data (image signals of moving images). It is a standard. MPE
The transmission speed for G1 standard is mainly about 1.5Mbps
Since it is limited to s, the moving image data is compressed to 2 to 15 Mbps in MPEG2 standardized to meet the demand for higher image quality.

Further, at present, MPEG1, MPEG2
And the working group that has promoted standardization (ISO / IEC JTC1 / SC2
9 / WG11), a moving image data compression technology that enables encoding processing and signal operation on an object basis and realizes new functions required in the multimedia age is being standardized as MPEG4. At first, MPEG4 aimed to standardize a low bit rate encoding method.
The standardization target has been extended to a more general-purpose encoding process of a high bit rate corresponding to an interlaced image.

FIG. 8 is a schematic diagram for explaining an encoding process for each object. FIG. 8A shows an image space Ts obtained from a color signal composed of a luminance signal and a color difference signal corresponding to an object constituting one image, and FIG. 8B shows a shape indicating the shape of the object. 5 shows an image space Ss obtained from a signal. A black portion in the image space Ss shown in FIG. 8B is an internal region of the object So, and a color image corresponding to pixels included in the object (intra-object pixels) is used to generate a fish image ( The object) To is expressed.

[0008] In MPEG4, an image signal corresponding to an object, including a shape signal and a color signal, is encoded in units of a region called a macroblock, which is composed of a predetermined number of pixels. The technology for encoding the color signal in units of a macroblock consisting of a predetermined number of pixels is known as MPEG.
1, introduced in MPEG2, but in MPEG4,
Along with the color signal, the shape signal is also coded on a macroblock basis.

FIG. 8 (c) is a diagram conceptually showing a block processing for dividing a color signal so as to correspond to the macroblock Tmb which partitions the image space Ts.
It is a figure which shows notionally the blocking process which divides a shape signal so that it may correspond to the macroblock Smb which divides the said image space Ss.

In MPEG4, an image signal corresponding to the entire screen is encoded instead of encoding an image signal corresponding to the entire screen so that encoding can be efficiently performed when an object is included in a part of the screen (image space). It is possible to encode only the image signal corresponding to the inside of the included rectangular area. In addition,
The rectangular area surrounding the object is called a bounding box.

Specifically, for a color signal, FIG.
As shown in (e), the block corresponding to the bounding box Tb is divided so as to correspond to the macroblock. Also, as shown in FIG. 8 (f), the shape signal is subjected to a block forming process of dividing a portion corresponding to the bounding box Sb into a macroblock.

An image signal including a color signal and a shape signal corresponding to each of the macroblocks is subjected to the same motion compensation coding processing as in MPEG1 and MPEG2 on a macroblock basis.

In the motion compensation coding process, the motion of an object between frames should originally be the same between the color signal and the shape signal. Does not coincide with the motion of the object for which the encoding processing of each signal is efficient.

That is, the motion compensation coding process is a method of coding a motion compensation error of a color signal or a shape signal. For this reason, in the motion compensation coding process, if motion information in which the motion compensation error is smaller than the motion information (motion vector) indicating the true motion of the object is obtained, the motion information (motion) indicating the true motion of the object is obtained. The motion compensation based on the motion information (motion vector) in which the motion compensation error is small leads to the improvement of the coding efficiency, rather than the motion compensation based on the motion vector. In particular, the color signal is a multi-valued signal, and the shape signal is a signal for distinguishing between the pixel inside the object and the pixel outside the object.
Since the signals are value signals, the properties of the two signals are slightly different, and the motion vector for reducing the motion compensation error is slightly different between the two signals.

[0015] The motion compensation error is calculated based on the color signal or shape signal corresponding to the macroblock to be processed (the macroblock to be coded) in the current screen (the screen to be coded), This is a difference value between a color signal or a shape signal corresponding to a predetermined region (prediction macro block) having the same size as the macroblock to be processed in a screen (an encoded reference screen). Also,
The motion vector corresponding to each macroblock is position information indicating the positional relationship between the macroblock to be processed and the predicted macroblock on the screen.

Therefore, in MPEG4, the motion compensation coding process for a color signal is performed by a motion vector obtained from the color signal (hereinafter, also referred to as a color motion vector).

[0017] Based on the above, the motion compensation coding process for the shape signal is performed based on a motion vector (hereinafter, also referred to as a shape motion vector) obtained from the shape signal.

The color motion vector and the corresponding shape motion vector do not always have the same direction and magnitude, but there is a strong correlation between these motion vectors. For this reason, in MPEG4, the shape motion vector corresponding to the macroblock to be encoded is predicted based on the shape motion vector corresponding to the encoded macroblock on the current screen and the corresponding color motion vector. A value is generated, and a difference value between the predicted value and the shape motion vector corresponding to the processed macroblock is encoded as information of the shape motion vector corresponding to the processed macroblock.

A description will now be given of a conventional image encoding apparatus and image decoding apparatus conforming to MPEG4. FIG.
FIG. 3 is a block diagram for explaining a conventional image encoding device conforming to PEG4. The image encoding apparatus 1000 includes a color encoding apparatus 1100 that performs an encoding process including a motion compensation encoding process on a color signal, and a shape encoding that performs an encoding process including a motion compensation encoding process on a shape signal. Device 1200.

The shape encoding apparatus 1200 includes a block generator 1210 that blocks the shape signal Ss corresponding to each object input to the input terminal 1201 so as to correspond to a macroblock, and outputs a blocked shape signal Bs. ,
A memory 1260 in which a shape signal (locally decoded shape data) Lds corresponding to the encoded screen (previous screen) is stored as a reference shape signal MLds.

Further, the shape encoding device 1200 operates to generate the reference shape signal MLd stored in the memory 1260.
s and a motion vector MVs corresponding to the macroblock to be processed, based on the block shape signal Bs output from the blocker 1210, and a motion compensation error signal related to a shape signal obtained at the time of the detection process. Based on the shape motion detector 1240 that outputs DFs, and the reference shape signal MLds and the shape motion vector MVs, the motion compensation process of the blocked shape signal Bs is performed, and the predicted shape signal (prediction) corresponding to the macroblock to be processed is obtained. A shape motion compensator 1250 that outputs a macroblock shape signal) Ps, and an output B of the blocker 1210
and a mode determiner 1280 that determines the coding mode of the shape signal based on the s and the motion compensation error signal DFs from the shape motion detector 1240, and outputs a shape coding mode signal Mos.

The shape encoding mode signal Mos
Is the macroblock to be processed corresponding to the shape signal, the macroblock outside the object, the intra macroblock subjected to the intra-screen encoding processing using the pixel value correlation in the screen, and the screen using the pixel value correlation between the screens This indicates which of the inter macroblocks has been subjected to the inter-coding process. The extra-object macroblock is a macroblock whose pixels are all located outside the object, and the intra-macroblock and the inter-macroblock are intra-object macroblocks in which at least some of the pixels are located inside the object.

Further, the shape encoding device 1200 performs an arithmetic encoding process (inter-frame coding) on the blocked shape signal Bs with reference to the predicted shape signal Ps based on the shape encoding mode signal Mos. Processing) or an arithmetic coding process (intra-screen coding process) that does not refer to the predicted shape signal Ps, to output the shape-coded data Cs, and to decode the shape-coded data Cs (local decoding Shape encoder 1220 that outputs Lds, and variable-length encoding that performs variable-length encoding on the shape-encoded data Cs and shape-encoding mode signal Mos and outputs a shape-encoded signal Es Vessel 1230
And information Mmv on the shape encoding mode signal Mos and the motion vector from the color encoding device 1100.
Based on t, Mmot, the shape motion vector MVs
And outputs a shape motion vector coded signal Emvs.

Also, in this shape encoding device 1200,
The locally decoded shape data Lds from the shape encoder 1220 is stored in the memory 1260 as a reference shape signal MLds.
It is stored as. This memory 126
The reference shape signal MLds stored in 0 is output to the shape motion detector 1240 and the shape motion compensator 1250, and is used for a motion detection process and a motion compensation process for the shape signal. The locally decoded shape data Lds is output to the color encoding device 1100, and is used for a motion detection process and a motion compensation process for a color signal.

On the other hand, the color encoding device 1100
A blocker 1110 that blocks a color signal St corresponding to each object input to the input terminal 1101 so as to correspond to a macroblock and outputs a block color signal Bt, and a coding-processed screen (previous screen). A memory 1160 in which a corresponding color signal (locally decoded color data) Ldt is stored as a reference color signal MLdt.

Further, the color encoding device 1100
The reference color signal MLd stored in the memory 1160
t, the locally decoded shape data Lds from the shape encoding device 1200, and the output B of the blocker 1110
It has a color motion detector 1140 that detects a color motion vector MVt corresponding to the macroblock to be processed based on t and outputs a motion compensation error signal DFt for the color signal.

The color encoding device 1100 is
Based on the output Bt of the blocker 1110, the motion compensation error signal DFt from the color motion detector 1140, and the blocked shape signal Bs from the shape encoder 1200, the coding mode for the color signal is determined. A mode determination unit 1180 that outputs a color encoding mode signal Mot, and a block color signal Bt based on the reference color signal MLdt, the color motion vector MVt, and the locally decoded shape data Lds from the shape encoding device 1200. Color motion compensator 11 that performs a motion compensation process of (1) and outputs a predicted color signal (color signal of the predicted macroblock) Pt for the processed macroblock.
50.

The color coding mode signal Mot
Are the macroblocks to be processed corresponding to the color signals, the macroblocks outside the object, the intra macroblocks subjected to the intra-frame encoding processing using the pixel value correlation in the screen, and the screen using the pixel value correlation between the screens This indicates which of the inter macroblocks has been subjected to the inter-coding process.

The color encoding device 1100 is
For the blocked color signal Bt, a waveform encoding process (inter-frame prediction encoding process) referring to the predicted color signal Pt or a waveform encoding process not referring to the predicted color signal Pt (intra-screen encoding process) To output the color coded data Ct, and decode data (locally decoded color data) L of the color coded data Ct.
and a color encoder 1120 that outputs dt.

Further, the color encoding apparatus 1100
A variable-length encoder 1130 that performs a variable-length encoding process on the color encoded data Ct and the color encoding mode signal Mot and outputs a color encoded signal Et;
Based on the color coding mode signal Mot, the color motion vector MVt is coded to output a color motion vector coded signal Emvt, and information Mmvt and Mmot relating to the color motion vector are converted to the motion vector of the shape coding device 1200. And a motion vector encoding device 1170 for outputting to the encoding device 1270.

In the color encoding apparatus 1100, the locally decoded color data Ldt from the color encoder 1120 is stored in the memory 1160 as a reference color signal MLdt. The stored reference color signal MLdt is output to the color motion detector 1140 and the color motion compensator 1150, and is used for a motion detection process and a motion compensation process.

Note that the image signal in MPEG4 has a data structure including a transparency signal indicating a combination ratio when a color signal corresponding to an object is combined with another color signal. Shape encoding device 1200
And a color coding device 1100, and a transparency coding device for coding the above-described transparency signal, but the description thereof is omitted here. It should be noted that this transmissivity encoding device has substantially the same configuration as the above-described color encoding device 1100. 1100. Therefore, the transparency encoder does not have a motion detector.

Next, the operation of the image coding apparatus 1000 will be described. When a shape signal Ss constituting an image signal corresponding to a predetermined object is input to the input terminal 1201 of the shape encoding device 1200, the shape signal Ss is converted by the blocking unit 1210 into the rectangular area (bounding box). Are divided so as to correspond to the respective macroblocks, and a blocked shape signal Bs is generated. The blocked shape signal Bs is output to the shape motion detector 1240, the shape encoder 1220, and the mode determiner 1280 in the shape encoder 1200, and is also output to the mode determiner 1180 of the color encoder 1100. You.

The shape motion detector 1240 detects a shape motion vector MVs corresponding to the macroblock to be processed, based on the block shape signal Bs and the reference shape signal MLds of the previous screen stored in the memory 1260. You. The shape motion vector MVs is output to the shape motion compensator 1250 and the motion vector coding device 1270.

At this time, the shape motion detector 12
From 40, the motion compensation error signal DFs related to the shape signal generated during the motion detection process is
Output to 0. Then, the mode determiner 1280 determines the coding mode for the macroblock to be processed based on the output Bs of the blocking unit 1210 and the motion compensation error signal DFs, and outputs the shape coding mode signal Mos. You.

The shape motion compensator 1250 performs a motion compensation process on the blocked shape signal Bs based on the shape motion vector MVs and the reference shape signal MLds, and obtains a predicted shape signal Ps corresponding to the macroblock to be processed.
Is generated. Further, the shape motion vector encoding device 1
In 270, information Mmvt, which relates to the shape encoding mode signal Mos and the motion vector from the motion vector encoding device 1170 in the color encoding device 1100,
The coding process of the shape motion vector MVs is performed based on the Mmot and the shape motion vector coded signal Em
vs is output from the output terminal 1203.

In the shape encoder 1220, the arithmetic coding process referring to the predicted shape signal Ps or the arithmetic coding process not referring to the predicted shape signal Ps is performed based on the shape coding mode signal Mos. This is performed on the blocked shape signal Bs to generate shape-encoded data Cs, and locally decoded shape data Lds is generated by arithmetic decoding of the shape-encoded data Cs. The locally decoded shape data Lds is stored in the memory 1
260 is stored as a reference shape signal MLds.

In the variable length encoder 1230, the shape coded data Cs and the shape coded mode signal Mo are generated.
A process of converting s into a variable length code is performed, and a shape coded signal Es corresponding to the macroblock to be processed generated by the variable length coding process is output from an output terminal 1202.

On the other hand, when a color signal St constituting an image signal corresponding to a predetermined object is input to the input terminal 1101 of the shape encoding device 1100, the color signal St is converted by the block converter 1110 into the rectangular area. A block color signal St is generated by being divided so as to correspond to each macroblock that divides the (bounding box).
This blocked color signal Bt is supplied to the color motion detector 1
140, color encoder 1120, and mode determiner 1
180 is output.

The color motion detector 1140 responds to the macroblock to be processed based on the block color signal Bt, the reference color signal MLdt in the memory 1160, and the locally decoded shape data Lds from the shape encoding device 1200. Is detected. The color motion vector MVt is output to the color motion compensator 1150 and the motion vector coding device 1170.

At this time, the color motion detector 1
From 140, the motion compensation error signal DFt for the color signal generated during the motion detection process is
180 is output. Then, the mode determiner 1180
In the above, based on the output Bt of the blocker 1110, the motion compensation error signal DFt, and the block shape signal Bs, the coding mode for the macroblock to be processed is determined, and the color coding mode signal Mot is output. You.

In the color motion compensator 1150, the color motion vector MVt and the reference color signal MLd
t, and the above-mentioned locally decoded shape data Lds,
The motion compensation of the blocked color signal Bt generates a predicted color signal Pt corresponding to the macroblock to be processed. Also, the color motion vector coding device 1170
Then, the encoding process of the color motion vector MVt is performed based on the color encoding mode signal Mot.
Thereby, the motion vector coded signal Emvt is output from the output terminal 1103, and the information Mmvt and Mmot regarding the color motion vector MVt are transmitted from the color motion vector coding device 1170 to the motion vector coding of the shape coding device 1200. Output to device 1270.

In the color encoder 1120,
On the basis of the color coding mode signal Mot and the locally decoded shape data Lds, a waveform coding process that refers to the predicted color signal Pt or a waveform code that does not refer to the predicted color signal Pt is performed on the blocked color signal Bt. The color coding data Ct is generated, and the locally decoded color data Ldt is generated by the waveform decoding processing of the color coded data Ct. The locally decoded color data Ldt is stored in the memory 1
160 stores the reference color signal MLdt.

The variable-length encoder 1130 converts the color-coded data Ct and the color-coding mode signal Mot into a variable-length code. Is output from the output terminal 1102.

FIG. 10 is a block diagram for explaining an image decoding apparatus conforming to MPEG4. This image decoding device 2000 decodes an image encoded signal output from the image encoding device 1000 shown in FIG.
A color decoding device 2100 that performs a decoding process including a motion compensation decoding process on the color encoded signal Et from the color encoding device 1100, and a shape encoded signal Es from the shape encoding device 1200. On the other hand, shape decoding device 22 that performs decoding processing including motion compensation decoding processing
00.

The shape decoding device 2200 performs variable-length decoding on the shape-encoded signal Es corresponding to each object input to the input terminal 2201, and performs shape-decoded data As and a shape-encoded mode signal. For the variable length decoder 2210 that outputs Mos and the shape decoded data As corresponding to the macroblock to be processed, refer to the predicted shape signal Ps of the macroblock to be processed based on the shape encoding mode signal Mos. Arithmetic decoding processing that does not refer to the predicted shape signal Ps,
And a shape decoder 2220 that outputs a shape decoded signal Ds.

The shape decoding device 2200 converts the shape decoded signal Ds into a decoded screen (previous screen).
2 for storing as a reference shape signal MDs corresponding to
260 and a rectangular area (bounding box) composed of a predetermined macroblock
And an inverse blocker 2230 that outputs the shape reproduction signal Rs from the output terminal 2203.

Further, the above-mentioned shape decoding device 2200
The image encoding device 1 input to the input terminal 2202
000 from the shape motion vector encoded signal Emvs.
Information Mmvs, M of the shape encoding mode signal Mos and the color motion vector from the color decoding device 2100
The motion vector decoding device 224 that decodes based on the mot to generate a decoded shape motion vector signal Dmvs.
0 and a shape motion compensator 22 that generates a predicted shape signal Ps corresponding to the macroblock to be processed based on the decoded shape motion vector signal Dmvs and the reference shape signal MDs.
50.

On the other hand, the color decoding device 2100
The color-coded signal Et corresponding to each object input to the input terminal 2101 is subjected to variable-length decoding to perform color-decoded data At and a color-coded mode signal Mot.
And the color encoding mode signal Mot and the shape decoding device 2 for the color decoded data At corresponding to the macroblock to be processed.
Based on the shape decoded signal Ds from the waveform decoding unit 200, a waveform decoding process that refers to the predicted color signal Pt of the macroblock to be processed or a waveform decoding process that does not refer to the predicted color signal Pt is performed. And a color decoder 2120 that outputs Dt.

The color decoding device 2100 is
A memory 2160 for storing the color decoded signal Dt as a reference shape signal MDt corresponding to a decoded screen (previous screen), and a rectangular area (bounding and ), And outputs a color reproduction signal Rt from an output terminal 2103.
And

Further, the color decoding device 2200
Is a color motion vector coded signal Emv input from the image coding device 1000 input to the input terminal 2102.
t is decoded based on the color coding mode signal Mot to generate a color motion vector decoded signal Dmvs and color motion vector information Mmvt, M
a motion vector decoding device 2140 that outputs mot;
A color motion compensator 21 that generates a predicted color signal Pt corresponding to a macroblock to be processed based on the color motion vector decoded signal Dmvt, the reference color signal MDt, and the shape decoded signal Ds from the shape decoding device 2200.
50.

Next, the operation of the image decoding apparatus 2000 will be described.

The input terminal 2 of the shape decoding device 2200
When a shape-encoded signal Es corresponding to each object is input to 201, the variable-length decoder 2210 performs a variable-length decoding process on the shape-encoded signal Es for each macroblock, Shape decoded data As and a shape encoding mode signal Mos corresponding to the macroblock are generated.

The input terminal 220 of the device 2200
2 is a shape motion vector coded signal Emvs
Is decoded by the motion vector decoding device 2240 based on the shape encoding mode signal Mos and the motion vector information Mmvs, Mmot from the color decoding device 2100 to generate a shape motion vector decoded signal Dmvs. Is done. Then, in the shape motion compensator 2250, the shape motion vector decoded signal Dmvs and the memory 22
60 based on the reference shape signal MDs stored in
A predicted shape signal Ps corresponding to the macroblock to be processed is generated.

In the shape decoder 2220, the shape decoded data As corresponding to the macroblock to be processed is calculated based on the shape encoding mode signal Mos.
An arithmetic decoding process that refers to the predicted shape signal Ps of the macroblock to be processed or an arithmetic decoding process that does not refer to the predicted shape signal Ps is performed to generate a shape decoded signal Ds. This shape decoded signal Ds is stored in the memory 2260.
As the reference shape signal MDs, and is output to the deblocker 2230. The deblocker 2
In 230, the shape-decoded signal Ds is integrated so as to correspond to a rectangular area (bounding box) composed of a predetermined macroblock, and is output from the output terminal 2203 as a shape reproduction signal Rs.

On the other hand, a color coded signal E corresponding to each object is input to an input terminal 2101 of the color decoding device 2100.
When t is input, the variable-length decoder 2110 performs a variable-length decoding process on the color-coded signal Et for each macroblock, and decodes the color-decoded data At and Color encoding mode signal Mo
t is generated.

The input terminal 210 of the device 2100
2 is a color motion vector coded signal Emvt input to
Are decoded by the motion vector decoding device 2140 based on the color coding mode signal Mot. The a decoding device 2140 outputs a color motion vector decoded signal Dmvt and also outputs information Mmvt and Mmot regarding the color motion vector. Then, in the color motion compensator 2150, the color motion vector decoded signal Dmvt, the reference color signal MDt stored in the memory 2160, and the shape decoder 22
The predicted color signal Pt corresponding to the macroblock to be processed is generated based on the shape decoded signal Ds from the macroblock 20.

In the color decoder 2120, the color decoded data A corresponding to the macroblock to be processed is output.
t, based on the color coding mode signal Mot and the shape decoded data Ds from the shape decoding device 2000, the predicted color signal Pt of the macroblock to be processed.
Or a waveform decoding process that does not refer to the predicted color signal Pt is performed to generate a color decoded signal Dt. This color decoded signal Dt
Are stored in the memory 2160 as the reference color signal MDt and output to the deblocker 2130.
In the deblocker 2130, the color decoded signal Dt
Are integrated so as to correspond to a rectangular area (bounding box) composed of predetermined macroblocks, and output from the output terminal 2103 as a color reproduction signal Rt.

Next, the encoding processing of the color motion vector and the shape motion vector in the above-described image encoding apparatus 1000 will be described in detail. First, a motion compensation process using a motion vector will be described. In MPEG4, in a frame unit motion compensation process for a color signal (interlaced or non-interlaced) forming one macroblock (image space composed of 16 × 16 pixels), basically, as shown in FIG. In addition, a predicted color signal corresponding to the macroblock MB to be processed is generated by the four motion vectors MV1 to MV4. Here, the four motion vectors MV1 to MV4 are shown in FIG.
As shown in (a), four blocks (image space composed of 8 × 8 pixels) B1 constituting one macroblock MB
Motion vectors corresponding to .about.B4. When all of these motion vectors MV1 to MV4 are equal, the motion compensation processing for the macroblock MB in frame units is performed as shown in FIG.
This is equivalent to a motion compensation process using V (MV = MV1 = MV2 = MV3 = MV4).

In the motion compensation process for each color signal (interlace) forming one macroblock in units of fields, as shown in FIG. 11B, the first motion vector MVf1 and the second motion vector MVf2 are used. , A predicted color signal corresponding to the processed macroblock MB in the frame is generated. Here, the first motion vector MVf1 is a motion vector corresponding to a first half macroblock (see FIG. 11C) MBf1 composed of pixels on a scanning line of an odd field forming a frame. The second motion vector MVf2 is a motion vector corresponding to a second half macroblock (see FIG. 11C) MBf2 composed of pixels on the scanning lines of the even fields forming the frame. Each half macroblock is an image space composed of 16 × 8 pixels.
In (b), it is shown as a square area for convenience of explanation. In addition, the first and second motion vectors MVf1 and MVf2 are equal, and the motion compensation processing on a field-by-field basis for the macroblock MB includes one motion vector MV (MV = MVf1 = MVf2)
This is equivalent to the frame-by-frame motion compensation processing.

On the other hand, in the frame-based motion compensation process for the shape signal, unlike the frame-based motion compensation process for the color signal, as shown in FIG. A predicted shape signal corresponding to the block MB is generated.

Next, the encoding processing of the motion vector will be described. In the encoding processing of the motion vector (color motion vector) of the color signal corresponding to the macroblock to be processed (16 × 16 pixels), the encoding is performed based on the motion vector of the processed reference macroblock positioned adjacent to the macroblock to be processed. The difference value between the predicted motion vector predicted in this way and the motion vector of the macroblock to be processed is encoded as information on the motion vector of the macroblock to be processed.

Hereinafter, the color motion vector prediction processing will be described in detail. (1) Motion vector prediction processing in frame-based motion compensation processing (1a) When the processing target macro block MBx has four motion vectors, as shown in FIG. The predicted value of the motion vector MVt0a of B0a is calculated using the blocks RB1a, RB2a, and RB2b adjacent to the block B0a as reference blocks.
a, MVt2a, and MVt2b. Also, the block B0 in the processed macro block MBx
The predicted value of the motion vector MVt0b of the block b is obtained by using the blocks B0a, RB2b, and RB3 adjacent to the block B0b as a reference block in the same manner as the predicted value of the motion vector MVt0a.
a, MVt2b, and MVt3. The predicted value of the motion vector MVt0c of the block B0c in the processed macroblock MBx is also the same as that of the motion vector MV.
Similarly to the predicted value of t0a, the blocks RB1b, B0a, and B0b adjacent to the block B0c are used as reference blocks, and the motion vectors MVt1b, MVt of these blocks are used.
0a, MVt0b. Further, the predicted value of the motion vector MVt0d of the block B0d in the processing target macroblock MBx is calculated using the blocks B0c, B0a, and B0b adjacent to the block B0d as the reference blocks.
c, MVt0a, and MVt0b.

(1b) Processing target macro block MB0 is 1
In other words, when there are two motion vectors MVt0, in other words, the motion vectors MVt0a to MVt0 of the blocks B0a to B0d in the macroblock MBx to be processed.
When d is equal to each other, as shown in FIG. 12 (b), the predicted value of the motion vector MVt0 of the macroblock MB0 to be processed is determined by using the blocks RB1, RB2, RB3 adjacent to the macroblock MB0 to be processed as reference blocks. , The motion vectors MVt1, MVt2, M
It is generated based on Vt3. In the description of the motion vector prediction process in the above-described frame-unit motion compensation process, the reference macroblock adjacent to the macroblock to be processed has four motion vectors, and the macroblock subjected to the frame-unit motion compensation process. In the case where the reference macroblock has only one motion vector, it is assumed that all four blocks constituting the reference macroblock have the same motion vector, A predicted value of the motion vector of the block is generated. When the reference macroblock is a macroblock having two motion vectors and subjected to a field-based motion compensation process, each block in the macroblock to be processed is determined based on an average value of the two motion vectors. A predicted value of the motion vector is generated.

(2) Motion Vector Prediction Process Performed in Field-Based Motion Compensation Process In this field-based motion compensation process, as shown in FIG. This is a process corresponding to a case where there are two motion vectors MVt01 and MVt02. The first motion vector MV
t01 is a motion vector corresponding to a first half macroblock (16 × 8 pixels) MBf1 composed of pixels on a scanning line of an odd field forming a frame. The second
Is a motion vector corresponding to the second half macroblock MBf2 (16 × 8 pixels) composed of pixels on the scanning lines of the even field forming the frame. In this case, as shown in FIG. 12C, the first and second motion vectors MV of the macroblock MBy to be processed.
The predicted values of tf1 and MVtf2 are generated based on the motion vectors MVt1, MVt2, and MVt3 of the blocks RB1, RB2, and RB3 adjacent to the macroblock MBy to be processed as reference blocks.

In the above description of the motion vector prediction processing in the field-based motion compensation processing, the reference macroblocks adjacent to the macroblock to be processed are all 4 macroblocks.
The figure shows a case in which a macroblock having two motion vectors and subjected to motion compensation processing in a frame unit is used. If the reference macroblock has only one motion vector, the four Assuming that the blocks have the same motion vector, the motion vectors MVt01 and Mv2 of the two half macroblocks MBf1 and MBf2 constituting the macroblock to be processed MBy
A predicted value of Vt02 is generated.

If the reference macroblock is a macroblock having two motion vectors corresponding to an odd field and an even field and subjected to a field-based motion compensation process, the macroblock to be processed MB
The predicted value of the motion vector MVt01 of the half macroblock MBf1 corresponding to the odd field that forms y is generated based on the motion vector of the half macroblock of the odd field that forms the reference macroblock. Further, in this case, the predicted value of the motion vector MV02 of the half macroblock MBf2 corresponding to the even field that constitutes the processed macroblock MBy is based on the motion vector of the half macroblock of the even field that constitutes the reference macroblock. Generated.

Next, the shape motion vector prediction process in the motion compensation process for the shape signal will be described in detail. The macroblock to be processed corresponding to the shape signal always has only one motion vector as described above.
In the encoding processing of the motion vector (shape motion vector) of the shape signal corresponding to the macroblock to be processed (16 × 16 pixels), the motion vector is calculated based on the motion vector of the processed reference macroblock positioned adjacent to the macroblock to be processed. A difference value between the predicted motion vector predicted in this way and the shape motion vector of the processed macroblock is encoded as information of the shape motion vector of the processed macroblock. However,
As described above, there is a strong correlation between the motion vector of the shape signal and the motion vector of the color signal. It is generated based on both the color motion vector of the block and the shape motion vector of the processed reference macroblock.

That is, the macroblock to be processed MBs0
The predicted value of the shape motion vector MVs0 corresponding to
2 (d), the macroblocks RMBs1 and RMBs around the processed macroblock MBs0 corresponding to the shape signal
2, MVs1, MVs2, and M of RMBs3
Not only Vs3 but also the blocks RB around the macroblock MB0 to be processed corresponding to the color signal shown in FIG.
, RB2, RB3 color motion vectors MVt1, M
It is generated with reference to Vt2 and MVt3. In the above-described color motion vector and shape motion vector prediction processing, the adjacent macroblock adjacent to the macroblock to be processed may be an intra macroblock (a macroblock that has been subjected to intra-frame encoding processing) or a macro outside the object. In the case of a block (a macroblock in which all pixels are located outside the object), since no motion vector exists for an intra macroblock or a macroblock outside the object, only an adjacent macroblock having a motion vector is used as a reference macroblock. , A process of generating a predicted value of a motion vector for the processed macroblock.

FIG. 13 is a block diagram for explaining a conventional motion vector encoding device constituting the image encoding device 1000. In this figure, a motion vector encoding device (hereinafter, referred to as a color motion vector encoding device) 117 constituting the color encoding device 1100 is shown.
0, and a motion vector encoding device (hereinafter, referred to as a shape motion vector encoding device) 1270 that forms the shape encoding device 1200 is shown. In the description of the motion vector encoding apparatus, the motion vector prediction process
The macroblock MB0 to be processed has only one color motion vector, and blocks R around the macroblock to be processed
A case where B1 to RB3 are used as reference blocks (see FIG. 12B) will be described as an example.

The above color motion vector coding device 117
0 encodes a motion vector corresponding to a non-interlaced color signal. The apparatus 1170 includes: an MV memory 102 for storing a motion vector MVt of a macroblock to be processed input to an input terminal 1; M storing the color encoding mode signal Mot input to the terminal 2
V effective memory 103. Here, the color encoding mode signal Mot is a color motion indicating whether a macroblock to be processed in an image space obtained from a color signal is one of an intra macroblock and an extra-object macroblock, or an inter macroblock. There is a vector valid signal. Therefore, by this color motion vector effective signal, a reference macroblock located adjacent to the macroblock to be processed in the image space is
It can be determined whether the block is one of an intra macroblock and a non-object macroblock or an inter macroblock.

Further, the color motion vector encoding device 1170, based on the color encoding mode signal Mot from the MV effective memory 103,
From the color motion vectors MVt1, MVt2, MVt3 corresponding to the reference blocks RB1, RB2, RB3 in the coded macro block stored in MV2.
An MV predictor 104 that generates a prediction value (prediction value of a color motion vector of a macroblock to be processed) Pmvt;
Referring to the MV prediction value Pmvt from the V predictor 104, the color motion vector MVt of the macroblock MB0 to be processed
And an MV encoder 105 that performs an encoding process of 0 and outputs an encoded signal Emvt of a color motion vector.

Further, the color motion vector encoding device 1170 includes a front-stage switch 100 provided between the input terminal 1 and the MV encoder 105, and an output terminal 10
Rear-stage switch 10 provided between V encoder 105
The switches 100 and 101 are controlled to open and close by the color encoding mode signal Mot.

On the other hand, the shape motion vector coding device 1
270 encodes a motion vector corresponding to a non-interlaced shape signal. The device 1270 includes an MV memory 202 for storing a shape motion vector MVs of a macroblock to be processed input to the input terminal 4; An MV effective memory 203 for storing the shape encoding mode signal Mos input to the input terminal 5. Here, the shape encoding mode signal Mos indicates whether the processed macroblock in the image space obtained from the shape signal is one of an intra macroblock and an extra-object macroblock,
Alternatively, it is a shape motion vector effective signal indicating whether the block is an inter macro block. Therefore, by the shape motion vector effective signal Mos, a reference macroblock located adjacent to the macroblock to be processed in the image space obtained from the shape signal is one of the intra macroblock and the extra-object macroblock. Or an inter-macro block can be determined.

The shape motion vector coding apparatus 1
270 is a shape motion vector MVs1, MVs2, MVs corresponding to an encoded macroblock stored in the MV memory 202, based on the color encoding mode signal Mot and the shape encoding mode signal Mos.
3, an MV predictor 20 that generates an MV prediction value (prediction value of the shape motion vector of the macroblock to be processed) Pmvs
4 and the MV prediction value Pmvs from the MV predictor 204 to perform an encoding process on the shape motion vector MVs0 of the macroblock MBs0 to be processed, and output an encoded signal Emvs of the shape motion vector. Device 205. Here, the MV predictor 204 is the MV memory 1 of the color motion vector encoding device 1170.
The half pixel precision color motion vector supplied from 02 is converted to a one pixel precision color motion vector, and a prediction process is performed using the one pixel precision color motion vector obtained by the conversion. This is because the color encoding device performs the motion compensation process using a half-pixel accuracy motion vector, while the shape encoding device performs the motion compensation process using a one-pixel accuracy motion vector.

Further, the shape motion vector encoding device 1270 includes a front-stage switch 200 provided between the input terminal 4 and the MV encoder 205, and an output terminal 11 and an MV
Post-switch 201 provided between the encoder and the encoder 205
The switches 200 and 201 are controlled to open and close by the shape encoding mode signal Mos.

Next, the operation will be described. The color motion vector encoding device 1170 encodes the motion vector MVt obtained from the color signal and outputs the encoded signal Emvt.
At 270, the motion vector MVs obtained from the shape signal
Is encoded, and the encoded signal Emvs is output.

That is, the color encoding mode signal Mot
Is input to the input terminal 2 as a motion vector valid signal, the motion vector valid signal Mot causes the preceding switch 100 and the subsequent switch 1 of the MV encoder 105 to operate.
01 is controlled to open and close. Specifically, the color motion vector M corresponding to the input macroblock MB0 to be processed is
If Vt0 is the one that needs to be encoded, both switches 100 and 101 are conducting, while
When the color motion vector MVt0 corresponding to the macroblock MB0 to be processed does not need to be coded, the switches 100 and 101 are turned off.

In other words, when the macroblock to be processed is an intra macroblock subjected to intra-frame encoding processing, or when the macroblock to be processed is a macroblock outside the object in which all pixels are located outside the object, When the encoding process for the motion vector is unnecessary, both switches 100 and 101 are turned off. on the other hand,
When the macroblock to be processed is an inter macroblock on which the motion compensation coding process (inter-frame coding process) has been performed, the switches 100 and 101 are turned on.

The color motion vector MVt corresponding to each macro block is temporarily stored in the MV memory 102. Specifically, the color motion vector (see FIG. 11 (a)) corresponding to the block constituting each macroblock is M
It is stored in the V memory 102. And the MV memory 10
2, the macroblock M to be processed shown in FIG.
The motion vectors MVt1 to MVt3 corresponding to the reference blocks MB1 to MB3 around B0 are supplied to the MV predictor 104. At this time, the color motion vector effective signal Mot
Are temporarily stored in the MV effective memory 103.

The MV predictor 104 is supplied with a color motion vector valid signal Mmot (Mot) corresponding to the macroblock to be processed from the MV valid memory 103. Then, in the MV predictor 104, the valid signal M
ot, the motion vectors MVt of the reference blocks RB1 to RB3 located adjacent to the macroblock to be processed.
1 to MVt3 to be encoded or not, that is, whether each of the reference blocks RB1 to RB3 belongs to one of an intra macroblock and a non-object macroblock, or belongs to an inter macroblock Is determined. Then, a predicted value (MV predicted value) Pmvt of the motion vector corresponding to the macroblock MB0 to be processed is generated according to the determination result.

For example, when the reference blocks RB1 to RB3 all belong to the inter macroblock, the MV predictor 104 calculates the motion vectors MVt1 to MV corresponding to the encoded reference blocks RB1 to RB3.
From t3, a predicted value (MV predicted value) Pmvt of the motion vector corresponding to the macroblock MB0 to be processed is generated. This MV prediction value Pmvt is output to MV encoder 105. Note that a motion vector of a reference block belonging to an intra macroblock or a non-object macroblock is not used for generating a prediction value.

The MV encoder 105 refers to the MV prediction value Pmvt and calculates a difference value (color difference motion vector) between the motion vector MVt (MVt0) of the macroblock to be processed and the prediction value Pmvt. It is encoded as a value corresponding to the color motion vector of the processing macro block, and an encoded signal Emvt of the color motion vector is generated. This color motion vector encoded signal Emvt
Is output via the rear-stage switch 101.

Similarly, when the shape encoding mode signal Mos, that is, the shape motion vector valid signal Mos indicating whether the shape motion vector of the macroblock to be processed is valid, is input to the input terminal 5, the shape motion The first switch 20 of the MV encoder 205 is controlled by the vector valid signal Mos.
0 and the subsequent stage switch 201 are opened and closed. Specifically, when the motion vector MVs0 corresponding to the input macroblock MBs0 to be processed needs to be encoded, both switches 200 and 201 are turned on. On the other hand, when the motion vector MVs0 corresponding to the macroblock MBs0 to be processed does not need to be coded, the switches 200 and 201 are turned off.

In other words, when the macroblock to be processed is an intra macroblock on which intra-frame coding processing has been performed, or when the macroblock to be processed is a macroblock outside the object in which all pixels are located outside the object, If the encoding process for the motion vector is unnecessary, both switches 200 and 201 are turned off. On the other hand, when the macroblock to be processed is an inter macroblock on which the motion compensation coding process (inter-frame coding process) of the shape signal has been performed, both switches 200 and 201 are turned on.

The shape motion vector MVs corresponding to each macro block is temporarily stored in the MV memory 202.
Then, from the MV memory 202, as shown in FIG.
Motion vectors MVs corresponding to reference macroblocks RMBs1 to RMBs3 around the processed macroblock MB0
1 to MVs3 are supplied to the MV predictor 204. Also,
At this time, the macroblock M to be processed shown in FIG.
The motion vectors MVt1 to MVt3 corresponding to the blocks MBt1 to MBt3 located adjacent to B0 are supplied from the MV memory 102 to the MV predictor 204. At this time, the shape motion vector effective signal Mos is stored in the MV effective memory 203.

The valid signal Mmos (Mos) of the shape motion vector corresponding to the macroblock to be processed is supplied from the MV valid memory 203 to the MV predictor 204,
From the MV effective memory 103, the color motion vector effective signal Mmot (Mo
t) is supplied. Then, in the MV predictor 204, based on the valid signal Mos, the reference macroblocks RMBs1 to RM positioned adjacent to the macroblock to be processed.
It is determined whether or not to encode the motion vectors MVs1 to MVs3 of Bs3, that is, whether or not each of the reference macroblocks is one of an intra macroblock and a non-object macroblock or an inter macroblock. Will be Further, based on the color motion vector effective signal Mmot, it is determined whether or not to encode the motion vectors MVt1 to MVt3 of the reference blocks RB1 to RB3 positioned adjacent to the macroblock to be processed, that is, each of the reference blocks It is determined whether RB1 to RB3 belong to one of the intra macroblock and the extra-object macroblock or belong to the inter macroblock. Then, a predicted value (MV predicted value) Pmvs of the motion vector corresponding to the macroblock MBs0 to be processed is generated according to the determination result.

For example, the reference macro blocks RMB1 to RMB1
When the MB3 and the reference blocks RB1 to RB3 are all inter-macroblocks, the MV predictor 204 generates the encoded reference macroblocks RMB1 to RMB3.
From the motion vectors MVs1 to MVs3 corresponding to the macroblocks MB1 to MVs3 and the motion vectors MVt1 to MVt3 corresponding to the encoded reference blocks RB1 to RB3.
Predicted value (MV predicted value) P of the motion vector corresponding to s0
mvs is generated. This MV prediction value Pmvs is output to MV encoder 205. Note that a motion vector of a reference macroblock that is an intra macroblock or an extra-object macroblock or a motion vector of a reference block belonging to an intra macroblock or an extra-object macroblock is not used for generating a prediction value.

In the MV encoder 205, the difference value (shape difference motion vector) between the motion vector MVs of the macroblock to be processed and the prediction value Pmvs is referred to by referring to the MV prediction value Pmvs. And a coded signal Emvs of the shape motion vector is generated. The encoded signal Emvs of the motion vector is output via the rear-stage switch 201.

FIG. 14 is a block diagram for explaining a conventional motion vector decoding device constituting the image decoding device 2000. In this figure, the color decoding device 2
The motion vector decoding device (hereinafter, referred to as a color motion vector decoding device) 2140 constituting the 100 and the motion vector decoding device (hereinafter, referred to as a shape motion vector decoding device) constituting the shape decoding device 2200 22.
40 is shown. Note that, in the description of the motion vector decoding device, as in the description of the motion vector decoding device, the motion vector prediction process is performed such that the macroblock MB0 to be processed has only one color motion vector,
A case where blocks RB1 to RB3 around the macroblock to be processed are set as reference blocks (see FIG. 12B) will be described as an example.

The above color motion vector decoding device 214
0 is the color motion vector encoding device 11 shown in FIG.
70, a color motion vector coded signal E output from
mvt is decoded.

The color motion vector decoding device 214
0 is a color motion vector coded signal Emvt input to the input terminal 13 and corresponding to the macroblock to be processed.
Is decoded to obtain a color motion vector decoded signal Dmvt.
And an MV memory 304 for temporarily storing a decoded signal Dmvt of the color motion vector. When decoding a motion vector corresponding to the macroblock MB0 to be processed, , The MV memory 3
From 04, the motion vectors MVt1 to MVt3 of the reference blocks RB1 to RB3 (see FIG. 12C) adjacent to the macroblock to be processed are output.

The color motion vector decoding device 2140 includes an MV effective memory 302 for temporarily storing a color encoding mode signal Mot corresponding to each macroblock input to the input terminal 6, and an MV effective memory 302. 3
02 based on the coding mode signal Mmot corresponding to the macroblock to be processed from the reference block RB
From the motion vectors MVt1 to MVt3 of the macroblocks 1 to RB3, the MV prediction value Pmv of the motion vector of the macroblock to be processed.
and an MV predictor 305 that generates t.

Further, the color motion vector decoding device 2140 is provided between the input terminal 13 and the MV decoder 303 at the former stage switch 300 and between the output terminal 20 and the MV decoder 303. Switch 3
01, and both switches 300 and 301 are controlled to open and close by the coding mode signal Mot.

The shape motion vector decoding device 2
240 is a motion vector encoding device 127 shown in FIG.
0, a coded signal Emv of a shape motion vector output from 0
s is to be decoded.

This shape motion vector decoding device 2240
Is an MV decoder 403 that decodes an encoded signal Emvs of the shape motion vector corresponding to the macroblock to be processed, which is input to the input terminal 14, and generates a decoded signal Dmvs of the shape motion vector. An MV memory 404 for temporarily storing the decoded signal Dmvs of the vector, and from the MV memory 404, reference macroblocks RMB1 to RMB3 adjacent to the macroblock to be processed are provided.
The motion vectors MVs1 to MVs3 (see FIG. 12D)
Is output.

The shape motion vector decoding device 2
240 is an MV that stores the coding mode signal Mos corresponding to the macroblock to be processed, which is input to the input terminal 7
A valid memory 402 and an encoding mode signal Mm corresponding to the macroblock to be processed from the MV valid memory 402
os and the motion vectors MVs1 to MVs3 of the processed reference macroblocks RMB1 to RMB3 and the processed reference blocks RB1 to RB3 based on the coding mode signal Mmot corresponding to the macroblock to be processed from the MV valid memory 302. Motion vectors MVt1 to MVt3
From the motion vector M of the processed macroblock MBs0
MV predictor 40 that generates an MV prediction value Pmvs of Vs0
5 is provided. Note that the MV predictor 405 converts a half-pixel accuracy color motion vector supplied from the MV memory 304 of the color motion vector decoding device 2140 into a one-pixel accuracy color motion vector, and is obtained by this conversion. The configuration is such that prediction processing is performed using a color motion vector with one-pixel accuracy. This is because the color decoding device performs the motion compensation process using a half-pixel accuracy motion vector, while the shape decoding device performs the motion compensation process using a one-pixel accuracy motion vector.

Further, the shape motion vector decoding device 2240 includes a pre-stage switch 400 provided between the input terminal 14 and the MV decoder 403, and an output terminal 21
A post-stage switch 40 provided between the V-decoder 403 and the
The two switches 400 and 401 are controlled to open and close by the coding mode signal Mos.

Next, the operation will be described. In the color motion vector decoding device 2140, the color motion vector coded signal Emvt output from the color motion vector coding device 1170 shown in FIG.
The decoded signal Dmvt is generated. Further, in the shape motion vector decoding device 2240, the coded signal Emvs of the shape motion vector output from the shape motion vector coding device 1270 shown in FIG. 13 is decoded, and the decoded signal Dmvs is generated. .

That is, the coding mode signal Mot decoded by the variable length decoder 2110 of the color decoding device 2100 is a motion vector valid signal indicating whether or not the motion vector of the macroblock to be processed is valid. Is input to the input terminal 6, the preceding-stage switch 300 and the subsequent-stage switch 301 of the MV decoder 303 are controlled to open and close by the encoding mode signal Mot. Specifically, when the input coded signal of the color motion vector corresponding to the input macroblock to be processed needs to be decoded, both switches 300 and 301 are turned on, while If the encoded signal of the motion vector does not need to be decoded, both switches 300 and 301 are turned off.

In other words, when the macroblock to be processed has been subjected to intra-frame encoding processing, or when the macroblock to be processed is a macroblock located outside the object and its motion vector does not need to be encoded. , Both switches are turned off, and when the macroblock to be processed has been subjected to the motion compensation coding process (inter-frame coding process), both switches are set to O.
It becomes N state.

The coded signal Emvt of the color motion vector input to the input terminal 13 is decoded by the MV decoder 303 with reference to the predicted value Pmvt for the color motion vector of the macroblock to be processed. A vector decoded signal Dmvt is generated. This decoded signal Dmvt is temporarily stored in MV memory 304.

In the decoding process for the color motion vector MVt0 of the macroblock MB0 to be processed, FIG.
The decoded signal Dmvt corresponding to the motion vectors MVt1, MVt2, MVt3 corresponding to the reference blocks RB1 to RB3 shown in (c) is input to the MV predictor 305. The coding mode signal Mot indicating whether the motion vector corresponding to each macroblock is valid is temporarily stored in the MV valid memory 302. And MV predictor 3
05, the motion vectors MVt1 to MVt1 of the reference block located adjacent to the macroblock to be processed based on the coding mode signal Mmot from the MV valid memory 302.
It is determined whether or not the MVt3 is coded, that is, whether or not each of the reference blocks belongs to one of the intra macroblock and the extra-object macroblock or the inter-macroblock. Will be According to this determination result, the macroblock M to be processed
A predicted value (MV predicted value) Pmvt of the motion vector MVt0 corresponding to B0 is generated. Note that the specific processing in the generation of the predicted value is the same as that in the color motion vector coding apparatus. And the MV prediction value Pmvt
Is output to the MV decoder 303.

On the other hand, the coding mode signal Mo decoded by the variable length decoder 2210 of the shape decoding device 2200
s is input to the input terminal 7, the encoding mode signal M
os, the front-stage switch 4 of the MV decoder 403
00 and the subsequent stage switch 401 are controlled to open and close. Specifically, when the input coded signal of the shape motion vector corresponding to the input macroblock to be processed needs to be decoded, the switches 400 and 401 are turned on. If the encoded signal of the motion vector does not need to be decoded, both switches 400 and 401 are turned off.

In other words, when the macroblock to be processed has been subjected to the intra-frame encoding process, or when the macroblock to be processed is a macroblock located outside the object and the encoding process for the motion vector is unnecessary. , The switches 400 and 401 are turned off. If the macroblock to be processed has been subjected to motion compensation coding (screen magic coding), the switches 400 and 401 are turned off. And 401 are turned on.

The coded signal Emvs of the shape motion vector input to the input terminal 14 is output to the MV decoder 403 by the MV decoder 403.
Decoding is performed with reference to the predicted value Pmvs for the shape motion vector of the macroblock to be processed, and a decoded signal Dmvs of the shape motion vector is generated. This decoded signal Dm
vs is temporarily stored in the MV memory 404.

In the encoding process for the shape motion vector MVs0 of the macroblock MBs0 to be processed, FIG.
Reference macroblocks RMB1, RMB2, R shown in (d)
Motion vector MVs1, MVs2, MV for MB3
The decoded signal Dmvs corresponding to s3, and FIG.
The decoded signals Dmvt corresponding to the motion vectors MVt1, MVt2, MVt3 for the reference blocks RB1, RB2, Rb3 shown in FIG. Further, the coding mode signal Mos indicating whether or not the shape motion vector corresponding to each macroblock is valid is temporarily stored in the MV valid memory 402. Then, the MV predictor 405 is supplied with the shape motion vector valid signal Mos corresponding to the macroblock to be processed from the MV valid memory 402, and outputs the color motion vector valid signal corresponding to the macroblock to be processed from the MV valid memory 302. A signal Mmot is provided.

Then, the MV predictor 405 determines whether or not the shape motion vectors MVs1 to MVs3 of the reference macroblock positioned adjacent to the macroblock to be processed are coded based on the valid signal Mmos. That is, it is determined whether each of the reference macroblocks is one of an intra macroblock and a non-object macroblock, or an inter macroblock. In addition, based on the coding mode signal Mmot from the MV valid memory 302, the motion vectors MVt1 to MVt1 of the reference block positioned adjacent to the macroblock to be processed.
It is determined whether or not the MVt3 is coded, that is, whether each of the reference blocks belongs to one of the intra macroblock and the extra-object macroblock or the inter-macroblock. Then, according to the determination result, the predicted value Pmvs of the motion vector corresponding to the macroblock MBs0 to be processed.
Is generated. The specific processing in this prediction value generation is the same as that in the motion vector encoding device. Then, the MV prediction value Pmvs is output to the MV decoder 403.

[0109]

An image encoding process corresponding to MPEG4 is performed by converting an image signal to be processed into an M signal.
Like the PEG2, it can be extended to an interlaced image signal, and in this image encoding process, it is also possible to perform a motion compensation process on a field basis for a color signal. Field-based motion compensation processing in MPEG4 is the same as in MPEG2.
In this motion compensation process, as shown in FIG. 12C, each macroblock in one screen (one frame) corresponds to the blocks MBf1 and MBf2 of the first and second fields constituting the macroblock. Motion vector MVt
01 and MVt02 are assigned. That is, there are two motion vectors for each macroblock.

However, in the motion compensation processing of MPEG4, the difference motion vector between the motion vector of the macroblock to be processed and the motion vector prediction value predicted from the motion vector of the neighboring macroblock is calculated as the information of the motion vector of the macroblock to be processed. A more efficient motion vector coding method than the motion vector coding in MPEG2 has been introduced.

FIG. 15 is a block diagram for explaining a conventional motion vector coding apparatus corresponding to an interlaced color signal u. This motion vector encoding device 117
0a is a frame MV memory 102a that stores a frame-based motion vector, a field MV memory 102b that stores a field-based motion vector, and a field MV converter 110a that converts a frame-based motion vector into a field-based motion vector. And a frame MV converter 110b that converts a field-based motion vector into a frame-based motion vector.

The motion vector encoding device 117
0a is an effective MV storing an encoding mode signal Mot indicating whether or not the motion vector of the processed macroblock is effective.
The memory 103 and the stored encoding mode signal Mmo
frame MV predictor 1 that predictively generates a predicted value of a motion vector in a frame unit from a motion vector corresponding to a coded reference macroblock in a frame unit based on t
04a, and a field MV predictor 104b that predictively generates a predicted value of a field-based motion vector from a motion vector of a coded reference macroblock in field units based on the stored coding mode signal Mmot. doing.

Here, the frame MV predictor 104a
Refers to the motion vector stored in the frame MV memory 102a if the motion vector of the encoded reference macroblock is a frame unit, and if the motion vector of the encoded reference macroblock is a field unit,
The configuration is such that a frame-based motion vector obtained by converting the field-based motion vector stored in the field MV memory 102b by the frame MV converter 110b is referred to. Field MV predictor 104b
Refers to the motion vector stored in the field MV memory 102b if the motion vector of the coded reference macroblock is a field unit, and if the motion vector of the coded reference macroblock is a frame unit,
The frame MV memory 102a refers to a field-based motion vector obtained by converting the frame-based motion vector stored in the frame MV memory 102a by the field MV converter 110a.

Further, the motion vector encoding device 11
70a refers to the MV prediction value output from the MV prediction unit 104a, and refers to the frame MV encoder 105a that performs the encoding process of the motion vector in a frame unit and the MV prediction value output from the MV prediction unit 104b. For reference, a field MV encoder 105b that performs a process of encoding a motion vector in a field unit is provided.

The motion vector encoding device 117
0a is an input node S1i connected to the input terminal 1, a first output node S1 connected to the input of the frame MV memory 102a and the input of the frame MV encoder 105a.
a, having a second output node S1b connected to the inputs of the field MV memory 102b and the field MV encoder 105b, and an open third output node S1c, based on the color coding mode signal Mot. hand,
There is a preceding switch 100 that connects the input node S1i to any of the first to third output nodes.
Also, the motion vector encoding device 1170a includes an output node S2o connected to the output terminal 10, a first input node S2a connected to the output of the frame MV encoder 105a, and an output node S2a connected to the output of the frame MV encoder 105a. It has a second input node S2b connected to the output and an open third input node S2c, and based on the color encoding mode signal Mot, connects the output node S2o to the first to third inputs. It has a post-stage switch 101 connected to any of the nodes.

Next, the operation will be described. The input terminal 1 has a color motion vector Mvt of the macroblock to be processed.
Is input to the input terminal 2, and the color motion vector effective signal Mot of the macroblock to be processed is input. Then, the front-stage switch 100 and the rear-stage switch 101 are controlled to be switched based on the color motion vector valid signal Mot.

That is, when the color motion vector MVt does not require the encoding process, the input node S1i of the preceding switch 100 is connected to the open output node S1c.
, And the output node S2o of the subsequent switch 101 is connected to its open input node S2c.

When the color motion vector MVt is a frame-based motion vector requiring encoding processing, the input node S1i of the preceding-stage switch 100 is
The output node S1a is connected to the output node S1a, and the output node S2o of the subsequent switch 101 is connected to the first input node S2a.

At this time, the motion vector MVt of the macroblock to be processed input to the input terminal 1 is temporarily M
It is stored in the V memory 102a. Further, field M
The V converter 110a converts the frame-based motion vector stored in the MV memory 102a into a field-based motion vector.

First, if the color motion vector MVt is a field-based motion vector requiring encoding, the input node S1i of the preceding switch 100 is connected to the second output node S1b, and the output of the subsequent switch 101 is output. Node S2o is connected to its second input node S2b.

At this time, the motion vector MVt of the macroblock to be processed input to the input terminal 1 is temporarily M
It is stored in the V memory 102b. Furthermore, the frame MV
The converter 110b converts the field-based motion vector stored in the MV memory 102b into a frame-based motion vector.

When the motion vector MVt of the macroblock to be processed is a frame-based motion vector, a predicted value of the frame-based motion vector is generated by the frame MV predictor 104a, and the frame MV encoder At 105a, the motion vector MVt of the macroblock to be processed is encoded with reference to the prediction value, and an encoded signal Emvt1 of the frame unit motion vector is generated.
This coded signal Emvt1 is generated as a motion vector coded signal Emvt via the subsequent-stage switch 101.

At this time, if the reference macroblock has been subjected to the frame motion compensation processing, the frame MV predictor 104a refers to the motion vector stored in the frame MV memory 102a, and the reference macroblock is referred to as the reference macroblock. If the field motion compensation processing has been performed, the motion vector for each field stored in the field MV memory 102b is converted to the frame MV converter 110b.
Is referred to a frame-based motion vector obtained by the conversion.

On the other hand, when the motion vector MVt of the macroblock to be processed is a field-based motion vector, a predicted value of the field-based motion vector is generated by the field MV predictor 104b, and the field MV is calculated.
In the encoder 105b, the motion vector MVt of the macroblock to be processed is encoded with reference to the prediction value, and an encoded signal Emvt2 of the field unit motion vector is generated. This encoded signal Emvt2 is transmitted to the rear-stage switch 10
1 is generated as a motion vector coded signal Emvt.

At this time, the field MV predictor 104b
Then, if the reference macro block has been subjected to the field motion compensation processing, the field MV memory 102b
Is referred to, and if the reference macroblock has been subjected to frame motion compensation processing, the motion vector for each frame stored in the frame MV memory 102b is converted to the field MV converter 110.
A field-based motion vector obtained by conversion in a is referred to.

FIG. 16 is a block diagram for explaining a conventional motion vector decoding apparatus corresponding to an interlaced color signal. This motion vector decoding device 2140
a is a motion vector encoding device 1170a shown in FIG.
To decode a motion vector coded signal Emvt output from, and output a motion vector decoded signal Dmvt.

The motion vector decoding device 2140a
Decodes the coded signal of the frame unit motion vector of the macroblock to be processed with reference to the predicted motion vector obtained from the motion vector of the decoded reference macroblock, and outputs a decoded signal Dmvt1 of the frame motion vector. A frame decoder 303a and a coded signal of a field unit motion vector of a processed macroblock are decoded with reference to a predicted motion vector obtained from a motion vector of a decoded reference macroblock to decode a field motion vector. And a field decoder 303b that outputs the signal Dmvt2.

The motion vector decoding device 2140a
Is a frame MV memory 304a storing a frame motion vector decoded signal Dmvt1, a field MV memory 304b storing a field motion vector decoded signal Dmvt2, and a frame unit motion vector decoded signal It has a field MV converter 310a for converting a decoded signal of a vector into a decoded signal of a motion vector and a frame MV converter 310b for converting a decoded signal of a motion vector in a field unit into a decoded signal of a motion vector in a frame unit.

The motion vector coding device 214
0a is an effective MV storing an encoding mode signal Mot indicating whether or not the motion vector of the processed macroblock is effective.
The memory 302 and the stored encoding mode signal Mmo
t, the predicted value P of the motion vector in frame units
A frame MV predictor 305a that predictively generates mvt1 from a decoded motion vector in frame units, and a predicted value Pmvt2 of a motion vector in field units is decoded in field units based on the stored coding mode signal Mot. Field MV for predictive generation from motion vector
And a predictor 305b.

Here, the frame MV predictor 305a
Refers to the motion vector stored in the frame MV memory 304a if the motion vector of the decoded reference macroblock is a frame unit, and to the field MV if the motion vector of the decoded reference macroblock is a field unit. The configuration is such that a frame-based motion vector obtained by converting the field-based motion vector stored in the memory 304b by the frame MV converter 310b is referred to. The field MV predictor 305b
If the motion vector of the decoded reference macroblock is a field unit, the motion vector stored in the field MV memory 304b is referred to, and if the motion vector of the decoded reference macroblock is a frame unit, the frame MV memory The configuration is such that the field-based motion vector obtained by converting the frame-based motion vector stored in 304a by the field MV converter 310a is referred to.

The motion vector decoding device 214
0a is the input node S3i connected to the input terminal 13,
A first output node S3a connected to the input of the frame MV decoder 303a, a second output node S3b connected to the input of the field MV decoder 303b,
And an open third output node S3c. Based on the coding mode signal Mot, the input node S3c
a switch 300 for connecting i to any of the first to third output nodes. Also, the motion vector decoding device 2140a includes an output node S4o connected to the output terminal 20, the frame MV decoder 303
a input node S4a connected to the output of the field MV decoder 303b.
And a third stage that connects the output node S4o to any of the first to third input nodes based on the encoding mode signal Mot. The switch 301 is provided.

Next, the operation will be described. Input terminal 13
Is supplied with a coded signal Emvt of the motion vector of the macroblock to be processed, and the input terminal 6 receives the motion vector effective signal Mot of the macroblock to be processed. Then, the front-stage switch 300 and the rear-stage switch 301 are controlled to be switched based on the motion vector valid signal Mot.

When the motion vector valid signal Mot is stored in the MV valid memory 302, the MV based on the valid signal Mmot stored in the MV valid memory 302 is used.
The frame MV estimator 305a determines whether the motion vectors MVt1, MVt2, and MVt3 of the macroblock referred to by the predictors 305a and 305b are coded and whether the motion vector is a frame-based motion vector or a field-based motion vector. And the field MV predictor 305b.

First, the motion vector coded signal Em
If vt does not require decoding, the input node S3i of the preceding switch 300 is connected to the open output node S3c, and the output node S4o of the subsequent switch 301 is connected to the open input node S4c. .

When the coded signal Emvt of the motion vector of the macroblock to be processed is a motion vector of a frame unit requiring decoding, the input node S3i of the preceding switch 300 is connected to the first output node S3a. Connected, and the output node S4o of the subsequent switch 301 is connected to its first input node S4a.

At this time, the coded signal Em of the motion vector of the macroblock to be processed input to the input terminal 13 is input.
The vt is decoded by the frame MV decoder 303a with reference to the predicted motion vector Pmvt1 of the macroblock to be processed, and a decoded motion vector signal Dmvt1 is generated. This decoded signal Dmvt1 is temporarily stored in MV memory 304a. Further, the field MV converter 310a converts the frame-based motion vector stored in the MV memory 304a into a field-based motion vector.

Further, in the frame MV predictor 305a,
A frame motion for generating a predicted motion vector of a motion vector of a processed macroblock by referring to a motion vector of a decoded reference macroblock adjacent to the processed macroblock based on a motion vector effective signal of the processed macroblock. A compensation process is performed. At this time, if the reference macroblock has been subjected to the frame motion compensation processing, the motion vector stored in the frame MV memory 304a is referred to, and the reference macroblock has been subjected to the field motion compensation processing. If, the frame-based motion vector obtained by converting the field-based motion vector stored in the field MV memory 304b by the frame MV converter 310b is referred to.

Next, when the encoded signal Emvt of the motion vector of the macroblock to be processed is a field-based motion vector requiring decoding processing, the input node S3i of the preceding switch 300 becomes the second output node S3b. , And the output node S4o of the subsequent switch 301 is connected to the second input node S4b.

At this time, the coded signal Em of the motion vector of the macroblock to be processed input to the input terminal 13
The vt is decoded by the field MV decoder 303b with reference to the predicted motion vector Pmvt2 of the macroblock to be processed, and a decoded signal Dmvt2 of the motion vector is generated. This decoded signal Dmvt2 is temporarily stored in MV memory 304b. Further, the field MV converter 310b converts the frame-based motion vector stored in the MV memory 304b into a field-based motion vector.

The field MV estimator 305b refers to the motion vector of the decoded reference macro block adjacent to the macro block to be processed based on the motion vector effective signal of the macro block to be processed, and A field motion compensation process for generating a predicted motion vector of the motion vector is performed. At this time, if the reference macroblock has been subjected to the field motion compensation processing, the motion vector stored in the field MV memory 304b is referred to, and the reference macroblock has been subjected to the frame motion compensation processing. If, the frame-based motion vector obtained by converting the field-based motion vector stored in the frame MV memory 304a by the frame MV converter 310b is referred to.

A wide range of applications can be considered for interlaced color signals used in ordinary TV broadcasting and non-interlaced color and shape signals used for the Internet and databases. The field of application is not expected to be very large.

The next-generation televisions, which have been talked about in recent years, are likely to adopt the non-interlaced system. Therefore, from a practical point of view, they are used for encoding interlaced signals whose use is not so large. In addition, it is not preferable that the configurations of the encoding device and the decoding device are complicated.

Therefore, in order to enable the encoding and decoding of the interlaced shape signal while suppressing the configuration of the encoding device and the decoding device from becoming complicated, the encoding or decoding of the interlaced color signal is performed. A method is conceivable in which a circuit configuration for decoding is combined with a circuit configuration for encoding or decoding of a non-interlaced shape signal.

However, if the non-interlace coding process is performed on the interlace shape signal, the motion compensation process of the interlace shape signal is also performed on a frame basis corresponding to the non-interlace. In this case, if the color signal is an interlaced signal, there are two types of motion vectors relating to the color signal: a field unit and a frame unit. It is difficult to encode a vector.

The present invention has been made to solve such a conventional problem. The encoding and decoding of a motion vector relating to a non-interlaced shape signal are performed by
An image encoding apparatus and an image decoding apparatus, an image encoding method, and an image encoding method capable of performing well in combination with encoding and decoding of a motion vector for an interlaced color signal with a minimum circuit configuration extension Data storage storing an image decoding method, and an image processing program for causing a computer to perform an encoding process or a decoding process of a motion vector related to an interlaced shape signal by the image encoding method or the image decoding method. The purpose is to obtain a medium.

[0146]

SUMMARY OF THE INVENTION An image processing apparatus according to the present invention (Claim 1) includes a color signal corresponding to each object included in a predetermined image space for displaying the object in color and the object. Receiving an interlaced image signal including a shape signal indicating the shape of the image space, and performing adaptive encoding processing including frame-based and field-based motion compensation encoding processing on the image signal by a predetermined number for partitioning the image space. An image coding apparatus that performs a motion compensation coding process on an interlaced color signal in a frame unit or a field unit for each macro block including pixels of A color motion vector encoding device for encoding a color motion vector corresponding to a macroblock based on a prediction value thereof; A shape motion vector corresponding to a macroblock to be processed is obtained from a color motion vector and a shape motion vector corresponding to a processed macroblock, for performing frame-based motion compensation coding on the shape signal. A shape motion vector coding device for coding based on the predicted value, wherein the color motion vector coding device converts a field-based color motion vector corresponding to a processed macroblock into a frame-based color motion vector. When the color motion vector corresponding to the processed macroblock is a field-based motion vector, the shape motion vector encoding device is configured to include a motion vector conversion unit. Of the predicted macroblock to the processed macroblock. Frame shape motion vector and is obtained by a structure having a shape motion vector predictor which generates based on the color motion vector of a frame unit outputted from the motion vector conversion means.

The image coding apparatus according to the present invention (claim 2) provides a color signal corresponding to each object included in a predetermined image space for displaying the object in color and a shape indicating the shape of the object. Receiving an interlaced image signal including a signal, and performing an adaptive encoding process including a motion compensation encoding process in a frame unit and a field unit on the image signal. An image processing apparatus that performs block-by-block processing, and performs frame-by-frame or field-by-frame motion compensation coding on an interlaced color signal, and performs frame-by-field or field-by-block color processing corresponding to a macroblock to be processed. A color motion vector encoding device that encodes a motion vector based on its predicted value; A shape motion vector encoding device that encodes a shape motion vector corresponding to a macroblock to be processed in frame units for performing frame-based motion compensation encoding processing, based on a predicted value thereof; When the color motion vector of the processed macroblock is a frame-based motion vector, the encoding apparatus calculates the predicted value of the shape motion vector corresponding to the macroblock to be processed by the color motion vector corresponding to the processed macroblock. And when the color motion vector of the processed macroblock is a motion vector in units of fields, the predicted value of the shape motion vector corresponding to the macroblock to be processed is generated as a processed macroblock. Generate based only on corresponding frame-based shape motion vectors It is obtained by a structure having Jo motion vector predictor.

The image processing apparatus according to the present invention (claim 3) can correspond to individual objects included in a predetermined image space.
Receiving an interlaced or non-interlaced image signal including a color signal for displaying the object in color and a shape signal indicating the shape of the object, and performing frame-based and field-based motion compensation coding on the image signal; An image coding apparatus for performing adaptive coding processing for each macroblock consisting of a predetermined number of pixels for partitioning the image space, wherein the frame-based or field-based motion compensation coding is performed on a color signal. A color motion vector encoding device that encodes a color motion vector corresponding to a macroblock to be processed in a frame unit or a field unit for performing processing based on its predicted value, and a non-interlaced shape signal. Corresponds to a macroblock to be processed in frame units for performing motion compensation coding processing in frame units A shape motion vector encoding device that encodes a shape motion vector based on a predicted value thereof, wherein the shape motion vector encoding device is configured to, when receiving the non-interlaced image signal as the image signal, A predicted value of a shape motion vector corresponding to a block is generated based on a color motion vector and a shape motion vector corresponding to a processed macroblock, and when an interlaced image signal is received as the image signal, the processed macroblock Is configured to include a shape motion vector predictor that generates a predicted value of a shape motion vector corresponding to a motion vector based only on a frame-based shape motion vector corresponding to a processed macroblock.

The image decoding apparatus according to the present invention (claim 4) provides a color signal corresponding to each object included in a predetermined image space for displaying the object in color and a shape indicating the shape of the object. Receiving an encoded image signal corresponding to an interlaced image signal including a signal, and performing an adaptive decoding process including a frame-based and a field-based motion-compensated decoding process on the image-encoded signal. An image decoding apparatus for performing a macroblock consisting of a predetermined number of pixels to be divided, for performing a motion compensation decoding process in a frame unit or a field unit on a coded signal corresponding to an interlaced color signal. A color motion that decodes a color motion vector corresponding to a macroblock to be processed in a frame unit or a field unit based on its predicted value. A vector decoding device, and a frame unit for performing a frame unit motion compensation decoding process on an encoded signal corresponding to an interlaced shape signal, a shape motion vector corresponding to a macroblock to be processed,
A shape motion vector decoding device that performs decoding based on a prediction value obtained from a color motion vector and a shape motion vector corresponding to the processed macroblock, and the color motion vector decoding device corresponds to the processed macroblock. And a motion vector converting means for converting a field-based color motion vector to a frame-based color motion vector. When the motion vector is a motion vector, the predicted value of the shape motion vector corresponding to the macroblock to be processed is converted into a frame-based shape motion vector corresponding to the processed macroblock and a frame-based color output from the motion vector conversion means. Generate based on motion vector It is obtained by a structure having a shape motion vector predictor.

An image processing apparatus according to the present invention (claim 5) is adapted to correspond to individual objects included in a predetermined image space.
Receiving an image coded signal corresponding to an interlaced image signal including a color signal for displaying the object in color and a shape signal indicating the shape of the object; An image decoding apparatus that performs adaptive decoding processing including compensation decoding processing for each macroblock composed of a predetermined number of pixels that divides the image space, and performs an adaptive decoding process on an encoded signal corresponding to an interlaced color signal. A color motion vector that decodes a color motion vector corresponding to a macroblock to be processed in a frame unit or a field unit for performing a motion compensation decoding process in a frame unit or a field unit based on a predicted value thereof. A decoding device that performs frame-based motion compensation decoding on the encoded signal corresponding to the non-interlaced shape signal; And a shape motion vector decoding device that decodes a shape motion vector corresponding to a macroblock to be processed in a frame unit for performing processing on the basis of a predicted value of the shape motion vector. When the color motion vector of the processed macroblock is a frame-based motion vector, the predicted value of the shape motion vector corresponding to the macroblock to be processed is calculated based on the color motion vector and the shape motion vector corresponding to the processed macroblock. When the generated color motion vector of the processed macroblock is a motion vector of a field unit, the predicted value of the shape motion vector corresponding to the macroblock to be processed is calculated by the shape motion of the frame unit corresponding to the processed macroblock. Has a shape motion vector predictor that generates based only on vectors It is obtained by the configuration.

An image processing apparatus according to the present invention (claim 6) is capable of handling individual objects included in a predetermined image space.
Receiving an image coded signal corresponding to an interlaced or non-interlaced image signal including a color signal for displaying the object in color and a shape signal indicating the shape of the object; And an adaptive decoding process including a motion compensation decoding process in units of fields for each macroblock consisting of a predetermined number of pixels for partitioning the image space. A color motion vector that encodes a color motion vector corresponding to a macroblock to be processed in a frame unit or a field unit for performing a frame unit or a field unit motion compensation decoding process on a signal based on a predicted value thereof. A decryption device;
A shape for decoding a shape motion vector corresponding to a macroblock to be processed in a frame unit for performing a motion compensation decoding process in a frame unit for a coded signal of a non-interlaced shape signal based on a predicted value thereof A motion vector decoding device, wherein when the shape motion vector decoding device receives an image coded signal corresponding to a non-interlaced image signal as the image coded signal, a shape corresponding to the macroblock to be processed is provided. The predicted value of the motion vector is
Generated based on the color motion vector and the shape motion vector corresponding to the processed macroblock, and when the image coded signal corresponding to the interlaced image signal is received as the image coded signal, the image block corresponds to the macroblock to be processed. The configuration has a shape motion vector predictor that generates a predicted value of a shape motion vector based only on a shape motion vector in a frame unit corresponding to a processed macroblock.

According to the image processing method of the present invention (claim 7), an image signal corresponding to each object included in a predetermined image space is adaptively processed by a motion compensation coding process in a frame unit and a field unit. An image encoding method for encoding each macroblock consisting of a predetermined number of pixels that divides the image space by an encoding process, wherein an interlace shape signal indicating the shape of an object included in the image signal is generated. A shape motion vector encoding process for encoding a shape motion vector for performing a frame-based motion compensation encoding process for each macroblock based on a predicted value thereof, and color display of an object included in the image signal. A field-based color motion vector for performing a field-based motion compensation coding process on an interlaced color signal for A color motion vector conversion process of converting the predicted value of the shape motion vector corresponding to the macroblock to be processed into a shape motion vector corresponding to the processed macroblock. When the color motion vector corresponding to the processed macroblock is a motion vector in units of fields when generating with reference to the color motion vector, the color motion vector of the processed macroblock is converted by the color motion vector conversion processing. This refers to a frame-based color motion vector obtained by conversion.

An image processing method according to the present invention (claim 8) is an adaptive image processing method that includes an interlaced image signal corresponding to each object included in a predetermined image space, which includes a motion compensation coding process in frame units and field units. An image encoding method for encoding macroblocks each consisting of a predetermined number of pixels that divides the image space by a typical encoding process, the method comprising: displaying an object included in the image signal in color. A color motion vector code for coding a frame-based or field-based color motion vector for performing frame-based or field-based motion compensation coding on a signal for each macroblock based on its predicted value. Processing and an interlaced shape signal indicating the shape of the object included in the image signal in frame units. And a shape motion vector encoding process for encoding a shape motion vector for performing a compensation encoding process for each macroblock based on its predicted value. In the shape motion vector encoding process, When the color motion vector corresponding to the macroblock is a frame-based motion vector, the predicted value of the shape motion vector corresponding to the processed macroblock is referred to by referring to the shape motion vector and the color motion vector corresponding to the processed macroblock. When the color motion vector corresponding to the processed macroblock is a field-based motion vector, the predicted value of the shape motion vector corresponding to the processed macroblock is calculated by the shape motion vector corresponding to the processed macroblock. It is generated by referring only to the vector.

According to the image coding method of the present invention (claim 9), an interlaced or non-interlaced picture signal corresponding to each object included in a predetermined picture space is converted into a frame unit and a field unit by motion. An image coding method for coding each macroblock consisting of a predetermined number of pixels for partitioning the image space by adaptive coding processing including compensation coding processing, wherein an object included in the image signal is color-coded. Encode a frame-based or field-based color motion vector for performing frame-based or field-based motion compensation coding on a color signal to be displayed for each macroblock based on its predicted value Color motion vector encoding processing,
A shape motion vector for performing a frame-based motion compensation coding process on a non-interlaced shape signal indicating a shape of an object included in the image signal is encoded for each macroblock based on its predicted value. And a shape motion vector encoding process, wherein in the shape motion vector encoding process, when the image signal is a non-interlaced image signal, a predicted value of a shape motion vector corresponding to the macroblock to be processed is processed. Generated with reference to the shape motion vector and the color motion vector corresponding to the block, and when the image signal is an interlaced image signal, the predicted value of the shape motion vector corresponding to the macroblock to be processed is converted to the processed macroblock. It is generated by referring only to the corresponding shape motion vector.

According to the image decoding method of the present invention (claim 10), the image decoding method according to the present invention is characterized in that an image coded signal corresponding to an interlaced image signal corresponding to each object included in a predetermined image space is
An image processing method for decoding macroblocks each consisting of a predetermined number of pixels that partition the image space by adaptive decoding including frame-based and field-based motion compensation decoding. A shape motion vector for performing a frame-based motion compensation decoding process on an encoded signal corresponding to an interlaced shape signal indicating the shape of an object included in a macroblock is decoded for each macroblock based on its predicted value. Shape motion vector decoding process, and a field unit for performing a field unit motion compensation decoding process on a coded signal corresponding to an interlaced color signal for displaying an object included in the image signal in color. A color motion vector conversion process of converting the color motion vector of In the shape motion vector decoding process, when the predicted value of the shape motion vector corresponding to the processed macroblock is generated with reference to the shape motion vector and the color motion vector corresponding to the processed macroblock, the processed macroblock If the color motion vector corresponding to is a field-based motion vector, the frame-based color motion vector obtained by converting the color motion vector of the processed macroblock by the color motion vector conversion processing is referred to. is there.

According to the image decoding method of the present invention (claim 11), an image encoding signal for an interlaced image signal corresponding to an individual object included in a predetermined image space is subjected to motion compensation in frame units and field units. An image decoding method for decoding macroblocks each consisting of a predetermined number of pixels for partitioning the image space by adaptive decoding including decoding, wherein an object included in the image signal is displayed in color. A frame-based or field-based color motion vector for performing a frame-based or field-based motion compensation decoding process on an encoded signal corresponding to an interlaced color signal for performing
A color motion vector decoding process for decoding each macroblock based on the predicted value; and a frame-by-frame motion for an encoded signal corresponding to an interlaced shape signal indicating the shape of an object included in the image signal. The shape motion vector for performing the compensation decoding process includes a shape motion vector decoding process for decoding each of the macroblocks based on the prediction value thereof. In the shape motion vector decoding process, the processed macro block When the color motion vector corresponding to is a frame-based motion vector, a predicted value of the shape motion vector corresponding to the processed macroblock is generated with reference to the shape motion vector and the color motion vector corresponding to the processed macroblock. The color motion vector corresponding to the processed macroblock is When it is Le, and generates only with reference to the shape motion vector prediction value of the shape motion vector corresponding to the processed macro block, corresponding to the processed macroblock.

According to the image decoding method of the present invention (claim 12), an image encoding signal for an interlaced or non-interlaced image signal corresponding to an individual object included in a predetermined image space is converted into a frame unit and a field unit. An image decoding method for decoding macroblocks of a predetermined number of pixels that partition the image space by adaptive decoding including unit motion compensation decoding, which is included in the image signal. A frame-based or field-based color motion vector for performing frame-based or field-based motion-compensation decoding on an encoded signal corresponding to a color signal for displaying an object in color is used as a prediction value. A color motion vector decoding process for decoding each macroblock based on A shape to be decoded for each macroblock based on a predicted value of a shape motion vector for performing frame-based motion compensation decoding on an encoded signal corresponding to a non-interlaced shape signal indicating the shape of In the shape motion vector decoding process, when the image coded signal is an image coded signal corresponding to a non-interlaced image signal, the shape motion vector A prediction value is generated with reference to a shape motion vector and a color motion vector corresponding to the processed macroblock, and when the image-encoded signal is an image-encoded signal corresponding to an interlaced image signal, The predicted value of the shape motion vector corresponding to the It is intended to.

[0158] A data storage medium according to the present invention (claim 13) is a data storage medium storing an image processing program, wherein the image processing program executes the coding process according to the image coding method according to claim 7. It stores an encoding program to be executed by a computer.

[0159] A data storage medium according to the present invention (claim 14) is a data storage medium storing an image processing program. It stores an encoding program to be executed by a computer.

[0160] A data storage medium according to the present invention (claim 15) is a data storage medium storing an image processing program. It stores an encoding program to be executed by a computer.

[0161] A data storage medium according to the present invention (claim 16) is a data storage medium storing an image processing program.
A decoding program for causing a computer to perform a decoding process according to the described image decoding method is stored.

A data storage medium according to the present invention (claim 17) is a data storage medium storing an image processing program, wherein the image processing program is used as the data storage medium.
A decoding program for causing a computer to perform a decoding process according to the described image decoding method is stored.

A data storage medium according to the present invention (claim 18) is a data storage medium storing an image processing program.
A decoding program for causing a computer to perform a decoding process according to the described image decoding method is stored.

[0164]

Embodiments of the present invention will be described below. (Embodiment 1) FIG. 1 is a block diagram for explaining a motion vector encoding apparatus constituting an image encoding apparatus according to Embodiment 1 of the present invention. The image encoding apparatus according to the first embodiment is an image encoding apparatus that encodes an interlaced image signal conforming to MPEG4, and includes a motion vector code of a color signal in a conventional image encoding apparatus 1000 shown in FIG. A motion vector coding device 1071 for the interlaced image signal is provided in place of the coding device 1170 and the shape signal motion vector coding device 1270.

The interlaced image signal conforming to the MPEG4 is an interlaced color signal for displaying an object forming one screen image in color, an interlaced shape signal indicating the shape of the object, and an interlaced shape signal indicating the shape of the object. The data structure includes an interlace transmittance signal indicating a combination ratio for combining with another image. Therefore, the image encoding device includes a color encoding device that performs the encoding process of the interlaced color signal, a shape encoding device that performs the encoding process of the interlaced shape signal, and a code of the interlace transparency signal. It has a transparency encoding device for performing a conversion process. However, the transparency encoding device has the same configuration as the color encoding device, and is not directly related to the present invention, so that the description thereof is omitted.

The motion vector coding apparatus 1701 according to the first embodiment includes a color motion vector coding apparatus 1171 constituting the color coding apparatus and a shape motion vector coding apparatus 1271 forming the shape coding apparatus. have. In the first embodiment, the color encoding apparatus has substantially the same configuration as the color encoding apparatus 1100 shown in FIG. 9, and performs a motion compensation encoding process on an interlaced color signal. The configuration is such that the unit-based motion compensation processing and the field-based motion compensation processing are adaptively switched. The color motion vector coding device 1171 in this color coding device
A conventional color motion vector encoding device 11 shown in FIG.
It has exactly the same configuration as 70a.

In the conventional explanation, the field converter 1
Although the specific processing in the color motion vector encoding device 11a and the frame converter 110b is not shown,
Specifically, the 71 field MV converter 110a includes:
The frame MV memory 102a receives a frame-based color motion vector (hereinafter, also referred to as a frame CMV) corresponding to a macroblock on which frame-based motion compensation processing has been performed, and receives a color corresponding to each field constituting the frame. Motion vector (hereinafter, field CMV)
Also say. ), The same two motion vectors as the frame CMV are output. Also, the frame M of the color motion vector encoding device 1171
Specifically, the V converter 110b reads two field-based color motion vectors (hereinafter, also referred to as a field CMV1 and a field CMV2) corresponding to a macroblock subjected to the field-based motion compensation processing from the field MV memory 102b. .), A single motion vector obtained by averaging the field CMV1 and the field CMV2 is output as a color motion vector (frame CMV) corresponding to the frame formed by the field.

The specific processing example of the frame MV converter 110b is not limited to the case where the average value of the motion vectors of both fields is used as the motion vector in the unit of frame. Is located outside the object, the motion vector of the half macroblock corresponding to the other field (that is, the half macroblock positioned inside the object) may be set as the motion vector of the macroblock corresponding to the frame. . In this case, the process of averaging the motion vector corresponding to each field becomes unnecessary, and the referenced motion vector is not affected by the motion vector having an insignificant value of the outer half macroblock. Has an effect.
In particular, when the output of the frame MV converter 110b can be used commonly by the frame MV predictor 104a in the color motion vector encoder 1171 and the MV predictor 204 in the shape motion vector encoder 1271, the motion vector averaging Therefore, there is a great effect in terms of reducing signal processing and improving prediction accuracy.

Also, in the first embodiment, the shape encoding device is the same as the shape motion vector encoding device 1271.
Of the conventional shape coding apparatus 1200 shown in FIG.
Is different. The shape encoding device according to the first embodiment is configured to convert an interlaced shape signal included in the interlaced image signal into a progressive shape signal, and to perform an encoding process on the progressive shape signal. That is, the shape coding apparatus is configured to always perform the motion compensation coding process in frame units during the inter-picture coding process.

Further, the shape motion vector coding device 1271 replaces the MV predictor 204 in the conventional shape motion vector coding device 1270 shown in FIG. 13 with the frame MV memory 102a in the color motion vector coding device 1171. Generates a prediction value for the shape motion vector of the macroblock to be processed with reference to the frame-based color motion vector stored in The MV predictor 204a is provided. Note that the MV predictor 204a uses the color motion vector encoding device 117
A configuration in which a half-pixel precision color motion vector supplied from one frame MV memory 102a is converted into a one-pixel precision color motion vector, and a prediction process is performed using the one-pixel precision color motion vector obtained by the conversion. It has become. This is because, in a color encoding device, a motion compensation process is performed using a half-pixel accuracy motion vector,
This is because the shape encoding device performs the motion compensation process using a motion vector with one-pixel accuracy.

More specifically, the MV predictor 204a generates a predicted value for the shape motion vector MVs0 of the macroblock MBs0 to be processed shown in FIG. Macroblock RMBs positioned adjacent to the macroblock MBs0 to be processed as a reference macroblock in the shape image space)
1 to RMBs3 are selected. Therefore, as the shape motion vector to be referred to, these macro blocks RMB
The motion vectors MVs1 to MVs3 of s1 to RMBs3 are used. On the other hand, in the image space (color image space) obtained from the color signal, the macroblock M
The reference macroblock located adjacent to the macroblock MB0 corresponding to Bs0 is not necessarily the one subjected to the motion compensation processing on a frame basis. Therefore, when the reference macroblock has been subjected to the motion compensation processing in units of frames, the motion vector stored in the frame MV memory 102a is referred to, and the reference macroblock is subjected to the motion compensation processing in units of fields. If the motion vector has been applied, the frame-based motion vector obtained by converting the field-based motion vector stored in the field MV memory 102b by the frame MV converter 110b is referred to.

It should be noted that whether or not the motion vectors MVt1, MVt2, MVt3 of each reference block in the color image space have been coded, in other words, the reference macroblock to which the reference block belongs has been subjected to the inter-frame coding process. The determination as to whether or not the block is an inter macro block is made in the MV predictor 204a based on the motion vector valid signal Mmot output from the MV valid memory 103.

Further, of the adjacent macroblocks RMBs1 to RMBs3 located adjacent to the macroblock MBs0 to be processed in the shape image space, those corresponding to the intra macroblock or the extra-object macroblock have their motion vectors. Are not referred to in the process of generating the predicted value of the motion vector. Similarly, among the adjacent blocks RB1 to RB3 positioned adjacent to the macroblock MB0 to be processed in the color image space, the motion vector of the adjacent block RB1 to RB3 belonging to the intra macroblock or the extra-object macroblock is also the motion vector. It is not referred to in the generation processing of the predicted value of the vector.

Next, the function and effect will be described. In the first embodiment, the operation of the image encoding device differs from the conventional image encoding device 1000 only in the operation of the motion vector encoding device 1071, and the operation of the motion vector encoding device 1071 has the shape Motion vector coding device 127
Only the operation 1 is different from the conventional motion vector coding apparatus. Therefore, the color motion vector encoding device 1 constituting the motion vector encoding device 1071 of the first embodiment
171 is a conventional color motion vector encoding device 1170
Is the same as

The MV predictor 204 in the conventional shape motion vector coding apparatus 1270 shown in FIG. 13 generates a color motion vector code as a color motion vector when generating a predicted value of the shape motion vector of the macroblock to be processed. Motion vectors MVt1, MVt2, MVt3 stored in the MV memory 102 in the conversion device 1170
Is referred to.

However, in the pixel coding apparatus corresponding to the interlaced color signal, the inter-frame coding processing is adaptively switched between the frame-based motion vector compensation coding processing and the field-based motion compensation coding processing. Therefore, two types of motion vectors, ie, a field-based motion vector and a frame-based motion vector, are generated as motion vectors corresponding to each macroblock.

In this case, since the shape motion vector referred to by the MV predictor 204 in the conventional shape motion vector encoding device 1270 is a frame-based motion vector, the MV predictor 204 uses a field-based color It is not possible to simply refer to a motion vector.

On the other hand, MV predictor 204a in shape motion vector coding apparatus 1271 of the first embodiment.
Then, when generating a predicted value of the shape motion vector of the macroblock to be processed in the shape image space, if the reference macroblock in the color image space has been subjected to motion compensation processing in units of frames, The corresponding motion vector stored in the frame MV memory 102a in the vector encoding device 1171 is referred to. On the other hand, if the reference macroblock in the color image space has been subjected to the motion compensation processing in units of fields, the MV predictor 204a outputs the field MV memory 1 in the color motion compensation encoding device 1171.
The frame-based motion vector obtained by converting the corresponding motion vector stored in 02b in the frame MV converter 110b is referred to.

Therefore, the shape motion vector coding device 12
In the 71 MV predictor 204a, the reference color motion vector is always a frame-based motion vector, regardless of whether the reference macroblock has been subjected to the field unit or frame motion compensation processing.

As a result, the motion vector encoding device 1171 in the image encoding device that performs the motion compensation encoding process on the interlaced image signal performs the motion compensation encoding process on the conventional non-interlaced image signal. Similar to the motion vector encoding device in the image encoding device, a prediction value of the shape motion vector corresponding to the processed macroblock in the shape image space is generated by referring to the color motion vector corresponding to the macroblock in the color image space. can do.

As described above, in the first embodiment, as the motion compensation coding process for an interlaced image signal including a color signal and a shape signal corresponding to an object, frame-based motion compensation coding and field-based motion compensation An image encoding apparatus that adaptively switches between encoding processing and MV prediction includes an MV predictor 204a that generates a predicted value of a shape motion vector with reference to a color motion vector.
In the predictor 204a, when the color motion vector to be referred to corresponds to the field unit motion compensation coding process, the frame MV converter 110b that converts the field unit color motion vector into a frame unit motion vector is used. Since the output is referred to, it is possible to refer to the color motion vector corresponding to the interlaced color signal also in the prediction processing of the shape motion vector in the shape coding device in which the motion compensation coding process is always performed in frame units. it can.

Therefore, the configuration of the shape motion vector coding apparatus is the same as that of the color motion vector coding apparatus, that is, the configuration having a motion vector coding unit corresponding to each of a frame-based motion vector and a field-based motion vector. By slightly changing the configuration of the MV estimator in the conventional shape motion vector coding device, the shape coding A simple circuit configuration of a shape motion vector coding apparatus that generates a predicted value of a shape motion vector with reference to a color motion vector when coding a vector can be realized.

In the first embodiment, the reference macroblock in the color image space referred to when generating the predicted value of the shape motion vector has one motion vector, No distinction is made between a case having a motion vector and a case having a motion vector, because the same motion vector is used in any of the above cases.

That is, when the reference macroblock has one motion vector, it is considered that the motion vectors corresponding to the four blocks constituting the reference macroblock are all equal, and thus the reference macroblock is eventually referred to. In any of the above cases, the motion vector of the processed macroblock MBs0 in the color image space (see FIG. 12C) is compared with the processed macroblock MBs0 in the shape image space (see FIG. 12D). The motion vectors MVt1 to MVt3 of the blocks RB1 to RB3 located in the periphery are used.

(Embodiment 2) FIG. 2 is a block diagram for explaining a motion vector encoding apparatus constituting an image decoding apparatus according to Embodiment 2 of the present invention. The image decoding apparatus according to the second embodiment is an image decoding apparatus that decodes an encoded signal obtained by encoding an interlaced image signal conforming to MPEG4, and a conventional image decoding apparatus shown in FIG. 2000, a motion vector decoding device 2041 for the interlaced image signal is provided instead of the motion vector decoding device 2140 for color signals and the motion vector decoding device 2240 for shape signals. In addition, the image decoding device includes a color decoding device that performs a decoding process corresponding to the interlaced color signal, a shape decoding device that performs a decoding process corresponding to the interlaced shape signal, and the interlace transmission device. Although it has a transparency decoding device that performs a decoding process corresponding to the degree signal, the transparency decoding device has the same configuration as the color decoding device, and is directly connected to the present invention. The description is omitted because it is not related.

The motion vector decoding device 2041 according to the second embodiment includes a color motion vector decoding device 2141 constituting the color decoding device, and a shape motion vector decoding device 2241 constituting the shape decoding device. have.

In the second embodiment, the color decoding device is the same as the color decoding device 210 shown in FIG.
0, and has a structure similar to that of 0. When performing motion compensation decoding processing on a coded signal obtained by coding an interlaced color signal, motion compensation processing in frame units and motion compensation processing in field units are applied. The configuration is such that the switching is performed dynamically. The color motion vector decoding device 2141 in this color decoding device has exactly the same configuration as the conventional color motion vector decoding device 2140a shown in FIG.

In the conventional description, the field converter 3
10a and the specific processing in the frame converter 310b are not shown, but the color motion vector decoding device 21
Specifically, the 41 field MV converter 310a includes:
The frame MV memory 304a receives a frame-by-frame color motion vector (frame CMV) corresponding to the macroblock subjected to the frame-by-frame motion compensation processing,
As a color motion vector (field CMV) corresponding to each field constituting the frame, the same two motion vectors as the frame CMV are output. The color motion vector decoding device 2
141, the frame MV converter 310b, specifically,
From the field MV memory 304b, two color motion vectors in the field unit corresponding to the macroblock subjected to the motion compensation processing in the field unit (field C
MV1 and field CMV2), and as a color motion vector (frame CMV) corresponding to a frame composed of two fields, the field CMV1
And a single motion vector obtained by averaging the field CMV2.

Further, a specific processing example of the frame MV converter 310b is not limited to a case where the average value of the motion vectors of both fields is set as a motion vector in frame units. Is located outside the object, the motion vector of the half macroblock corresponding to the other field (that is, the half macroblock positioned inside the object) may be set as the motion vector of the macroblock corresponding to the frame. . In this case, the process of averaging the motion vector corresponding to each field becomes unnecessary, and the referenced motion vector is not affected by the motion vector having an insignificant value of the outer half macroblock. Has an effect.
In particular, the output of the frame MV converter 310b is output from the frame MV predictor 305a in the color motion vector decoding device 2141 and the shape motion vector decoding device 2241.
If it can be used in common with the MV predictor 405a in
This is highly effective in reducing signal processing for motion vector averaging and improving prediction accuracy.

In the second embodiment, the shape decoding device is different from the shape motion vector decoding device 2241 in the second embodiment.
Is different from the shape decoding device 2200 shown in FIG. The shape decoding apparatus according to the second embodiment performs a decoding process on a shape-encoded signal obtained by encoding a progressive shape signal by the shape encoding apparatus according to the first embodiment. That is, the shape decoding apparatus is configured to always perform the motion compensation decoding process in units of frames during the inter-picture decoding process. Note that the progressive shape signal obtained by this decoding process is converted into an interlaced shape signal and used for image display.

Further, the shape motion vector decoding device 2141 replaces the MV predictor 405 in the shape motion vector decoding device 2240 constituting the conventional motion vector decoding device shown in FIG. MV memory 30 in the conversion device 2141
4a, the frame-based color motion vector stored in the frame MV converter 3
An MV predictor 405a that adaptively selects one of the frame-based color motion vectors output from 10b and generates a predicted value for the shape motion vector of the macroblock to be processed with reference to the selected color motion vector. Things.

More specifically, the MV predictor 405a
When generating a predicted value for the shape motion vector MVs0 of the processed macroblock MBs0 shown in FIG.
In the image space (shape image space) obtained from the shape signal,
The macroblock MB to be processed as a reference macroblock
macroblocks RMBs1 to RMB located adjacent to s0
MBs3 is selected. Therefore, these macro blocks RMBs1 to RMBs1
The motion vectors MVs1 to MVs3 of RMBs3 are used. On the other hand, in an image space (color image space) obtained from a color signal, a reference macroblock positioned adjacent to a macroblock MB0 corresponding to the macroblock MBs0 to be processed has been subjected to motion compensation processing in frame units. Is not always the case. Therefore, when the reference macroblock has been subjected to the motion compensation processing in units of frames, the motion vector stored in the frame MV memory 304a is referred to, and the reference macroblock is subjected to the motion compensation processing in units of fields. If the motion vector has been obtained, the frame-based motion vector obtained by converting the field-based motion vector stored in the field MV memory 304b by the frame MV converter 310b is referred to.

It should be noted that it is determined whether or not the motion vectors MVt1, MVt2, MVt3 of each reference block in the color image space are coded, in other words, whether or not the reference macroblock to which the reference block belongs is an inter macroblock. Is performed based on a motion vector valid signal Mmot output from the MV valid memory 402.

[0194] Of the adjacent macroblocks RMBs1 to RMBs3 located adjacent to the macroblock MBs0 to be processed in the above-mentioned shape image space, those corresponding to intra macroblocks or extra-object macroblocks have their motion vectors. Are not referred to in the process of generating the predicted value of the motion vector. Similarly, among the adjacent blocks RB1 to RB3 positioned adjacent to the macroblock MB0 to be processed in the color image space, the motion vector of the adjacent block RB1 to RB3 belonging to the intra macroblock or the extra-object macroblock is also the motion vector. It is not referred to in the generation processing of the predicted value of the vector.

Next, the function and effect will be described. In the second embodiment, the operation of the image decoding device differs from the conventional image decoding device 2000 only in the operation of the motion vector decoding device 2041, and the operation of the motion vector decoding device 2041 is Motion vector coding device 224
Only the operation 1 is different from the conventional motion vector decoding apparatus. Therefore, the color motion vector decoding device 2 constituting the motion vector decoding device 2041 of the second embodiment
141 is a conventional color motion vector decoding device 2140
Is the same as

The MV predictor 405 in the conventional shape motion vector decoding device 2240 shown in FIG. 14 generates a color motion vector decoding Motion vectors MVt1, MVt2, MVt3 stored in the MV memory 304 in the conversion device 2140
Is referred to.

However, in the pixel decoding apparatus corresponding to the interlaced color signal, the inter-frame decoding processing is adaptively switched between the frame-based motion vector compensation decoding processing and the field-based motion compensation decoding processing. Therefore, two types of motion vectors, a field-based motion vector and a frame-based motion vector, are generated as motion vectors corresponding to each macroblock.

In this case, since the shape motion vector referred to by the MV predictor 405 in the conventional shape motion vector decoding device 2240 is a frame-based motion vector, the MV predictor 405 uses the field-based color It is not possible to simply refer to a motion vector.

On the other hand, the MV predictor 405a in the shape motion vector decoding device 2241 according to the second embodiment.
In generating a predicted value of a shape motion vector of a macroblock to be processed in a shape image space, if the reference macroblock in the color image space has been subjected to motion compensation processing on a frame-by-frame basis, The corresponding motion vector stored in the frame MV memory 304a in the vector decoding device 2141 is referred to. On the other hand, if the reference macroblock in the color image space has been subjected to the motion compensation processing in units of fields, the MV predictor 405a uses the field MV memory 304b of the color motion compensation encoding device 2141.
The frame-based motion vector obtained by converting the corresponding field-based motion vector in the frame MV converter 310b is referred to.

Therefore, the shape motion vector decoding device 22
In the MV predictor 405a of 41, the color motion vector referred to is always a frame-based motion vector regardless of whether the reference macroblock has been subjected to motion compensation processing in field units or frames.

As a result, in the motion vector decoding device 2141 in the image decoding device which performs the motion compensation decoding process on the interlaced image signal, the motion compensated decoding is performed on the encoded signal of the conventional non-interlaced image signal. In the same manner as the motion vector decoding device in the image decoding device that performs the conversion process, the color motion vector corresponding to the macroblock in the color image space is referred to, and the shape motion vector corresponding to the processed macroblock in the shape image space is referred to. A predicted value can be generated.

As described above, in the second embodiment, the motion compensation decoding process for the encoded signal of the interlaced image signal including the color signal and the shape signal corresponding to the object is performed as follows.
An MV predictor that generates a predicted value of a shape motion vector by referring to a color motion vector in an image decoding device that adaptively switches between frame-based motion compensation decoding processing and field-based motion compensation decoding processing. The MV predictor 405a converts the field-based color motion vector into a frame-based motion vector when the color motion vector to be referenced corresponds to the field-based motion compensation decoding process. Since the output of the frame MV converter 310b is referred to, the color motion corresponding to the interlaced color signal can be performed in the shape motion vector prediction process in the shape decoding device in which the motion compensation decoding process is always performed in frame units. You can refer to a vector.

[0203] For this reason, the shape motion vector decoding device is configured to have a motion vector decoding unit corresponding to each of the frame-based motion vector and the field-based motion vector, similarly to the color motion vector decoding device. In addition, by slightly changing the configuration of the MV predictor in the conventional shape motion vector decoding device, an image decoding device that performs a motion compensation decoding process on an encoded signal of an interlaced image signal corresponding to an object has A simple circuit configuration of a shape motion vector decoding device that generates a predicted value of a shape motion vector by referring to a color motion vector when decoding a coded signal of a motion vector can be realized.

In the second embodiment, the case where the reference macroblock in the color image space referred to when generating the predicted value of the shape motion vector has one motion vector, and the case where four macroblocks corresponding to each block No distinction is made between a case having a motion vector and a case having a motion vector, because the same motion vector is used in any of the above cases.

That is, when the reference macroblock has one motion vector, it is considered that all the motion vectors corresponding to the four blocks constituting the reference macroblock are equal, and thus the reference macroblock is eventually referred to. In any of the above cases, the motion vector of the processed macroblock MBs0 in the color image space (see FIG. 12C) is compared with the processed macroblock MBs0 in the shape image space (see FIG. 12D). The motion vectors MVt1 to MVt3 of the blocks RB1 to RB3 located in the periphery are used.

(Embodiment 3) FIG. 3 is a block diagram for explaining a motion vector encoding apparatus constituting an image encoding apparatus according to Embodiment 3 of the present invention. The motion vector coding apparatus 1072 according to the third embodiment is similar to the motion vector coding apparatus 1071 according to the first embodiment, except that the color motion vector coding apparatus 1172 and the shape coding apparatus And a shape motion vector encoding device 1272 which constitutes

Here, the color motion vector coding device 1172 has exactly the same configuration as the color motion vector coding device 1171 of the first embodiment. Also, the shape motion vector encoding device 1272 is different from the shape motion vector encoding device 1271 of the first embodiment in that the MV predictor 204a in the color motion vector encoding device 1172
The MV predictor 204b includes a MV predictor 204b that generates a predicted value of a shape motion vector of a macroblock to be processed with reference to only the output of the macroblock 2a (that is, a motion vector for each frame). Therefore, the shape motion vector coding apparatus 1272 according to the second embodiment is different from the shape motion vector coding apparatus 1 according to the first embodiment in the configuration other than the MV predictor 204b.
271 is exactly the same.

More specifically, the MV predictor 204b generates a predicted value for the shape motion vector MVs0 of the macroblock MBs0 to be processed shown in FIG. Macroblock RMBs positioned adjacent to the macroblock MBs0 to be processed as a reference macroblock in the shape image space)
1 to RMBs3 are selected. Therefore, as the shape motion vector to be referred to, these macro blocks RMB
The motion vectors MVs1 to MVs3 of s1 to RMBs3 are used. On the other hand, in the image space (color image space) obtained from the color signal, the macroblock M
The reference macroblock located adjacent to the macroblock MB0 corresponding to Bs0 is not necessarily the one subjected to the motion compensation processing on a frame basis. Therefore, when the reference macroblock has been subjected to the motion compensation processing in units of frames, the motion vector stored in the frame MV memory 102a is referred to, and the reference macroblock is subjected to the motion compensation processing in units of fields. If so, the color motion vector is not referenced. The other configuration of the MV predictor 204b is the same as the MV predictor 204a of the first embodiment.

Next, the function and effect will be described. The motion vector coding apparatus 1072 according to the third embodiment differs from the motion vector coding apparatus 1071 according to the first embodiment in that
Frame MV of color motion vector coding device 1172
Output of converter 110b (frame-based motion vector)
Is not transmitted to the shape motion vector coding device 1272.

The shape motion vector encoding device 1
In 272, when generating a predicted value of the shape motion vector of the macroblock to be processed in the shape image space, if the reference macroblock in the color image space has been subjected to the motion compensation process in units of frames, Frame MV in motion vector coding apparatus 1171
The corresponding motion vector stored in the memory 102a is referred to by the MV predictor 204b. On the other hand, when the reference macroblock in the color image space has been subjected to the motion compensation processing on a field basis, the MV predictor 2
In 04b, unlike the MV predictor 204a in the first embodiment, the output (the motion vector in frame units) of the frame MV converter 110b in the color motion compensation encoding device 1171 is not referred to.

For this reason, the motion vector coding apparatus 1072 according to the third embodiment does not refer to the color motion vector on a field-by-field basis, and is therefore different from the motion vector coding apparatus 1071 according to the first embodiment (see FIG. 1). Therefore, there is a possibility that the prediction accuracy of the prediction value for the shape motion vector of the macroblock to be processed is deteriorated.

In the case where the macroblock to be processed has been subjected to motion compensation processing in units of frames and the reference macroblock has been subjected to motion compensation processing in units of fields, the color motion vector code Since the output of the frame MV converter 110b in the encoding device 1172 is referred to by the frame MV predictor 104a in the color motion vector encoding device 1172,
It can be said that the configuration of the motion vector encoding device according to the first embodiment is suitable.

That is, in the above case, the frame MV predictor 1 in the color motion vector coding device 1172
Since it is necessary to refer to the motion vector in field units at 04a, the frame MV converter 110b performs a process of converting the motion vector in field units into frame units. Therefore, the frame MV converter 110b
Is output to the frame MV predictor 10 of the color motion vector encoder 1172.
4a and the output to both the MV predictor 204b of the shape motion vector encoding device 1272 and the supply to one of the MV predictors 104a are the same in the amount of processing in the frame MV converter 110b. Therefore, it can be said that the configuration of the motion vector encoding apparatus according to the first embodiment is suitable for the above case from the viewpoint of the prediction accuracy of the predicted value of the shape motion vector.

However, for example, when both the macroblock to be processed and the reference macroblock in the color image space have been subjected to the motion compensation processing in units of fields, the color motion vector coding apparatus 1172
Then, the output of the field MV memory 304b is referred to. For this reason, in the configuration of the motion vector encoding device 1071 of the first embodiment that refers to the color motion vector of the reference macroblock on which the field-based motion compensation processing has been performed during the prediction processing of the shape motion vector, the shape motion vector In order to perform the prediction processing, the frame MV converter 110b must perform the processing of converting the field-based color motion vector into the frame-based color motion vector.

On the other hand, in the motion vector coding apparatus 1172 of the third embodiment, only the output of the frame MV memory 102a of the shape motion vector coding apparatus 1172 is
04b.

In such a configuration, for example, when both the macroblock to be processed and the reference macroblock in the color image space have been subjected to motion compensation processing in units of fields, the shape motion vector encoding device 12
In 72, the color motion vector of the reference macroblock is not referred to in the prediction process of the shape motion vector, and as a result, the color motion vector encoding device 11
The 72-frame MV converter 110b can omit a process of converting a field-based color motion vector into a frame-based color motion vector.

(Embodiment 4) FIG. 4 is a block diagram for explaining a motion vector decoding device constituting an image decoding device according to Embodiment 4 of the present invention. The motion vector decoding device 2042 of the fourth embodiment corresponds to the motion vector coding device 1072 of the third embodiment shown in FIG. This motion vector decoding device 204
2 is a color motion vector decoding device 2142 that decodes the color motion vector coded signal Emvt output from the color motion vector coding device 1172 in the motion vector coding device 1072 and outputs the decoded signal Dmvt. , The motion vector encoding device 10
72, a shape motion vector decoding device 2242 for decoding the shape motion vector encoded signal Emvs output from the shape motion vector coding device 1272 and outputting the decoded signal Dmvs.

The color motion vector decoding device 2142 has exactly the same configuration as the color motion vector decoding device 2141 of the second embodiment. Also, the shape motion vector decoding device 2142 includes a frame MV memory 30 in the color motion vector decoding device 2142 instead of the MV predictor 405a in the shape motion vector decoding device 2142 of the second embodiment.
An MV predictor 405b that generates a predicted value of a shape motion vector of a macroblock to be processed by referring only to the output of the macroblock 4a (that is, a motion vector in units of frames). Therefore, the configuration of the shape motion vector decoding device 2142 according to the fourth embodiment differs from the configuration of the shape motion vector decoding device 2 according to the second embodiment except for the configuration of the MV predictor 405b.
141 is exactly the same.

More specifically, the MV estimator 405b
When generating a predicted value for the shape motion vector MVs0 of the processed macroblock MBs0 shown in FIG.
In the image space (shape image space) obtained from the shape signal,
The macroblock MB to be processed as a reference macroblock
macroblocks RMBs1 to RMB located adjacent to s0
MBs3 is selected. Therefore, these macro blocks RMBs1 to RMBs1
The motion vectors MVs1 to MVs3 of RMBs3 are used. On the other hand, in an image space (color image space) obtained from a color signal, a reference macroblock positioned adjacent to a macroblock MB0 corresponding to the macroblock MBs0 to be processed has been subjected to motion compensation processing in frame units. Is not always the case. Therefore, when the reference macroblock has been subjected to the motion compensation processing in units of frames, the motion vector stored in the frame MV memory 304a is referred to, and the reference macroblock is subjected to the motion compensation processing in units of fields. If so, it does not refer to the color motion vector. Note that M
Other configurations of the V predictor 405b are the same as those of the second embodiment.
Is the same as the configuration of the MV predictor 405a.

Next, the function and effect will be described. The motion vector encoding device 2042 according to the fourth embodiment differs from the motion vector decoding device 2041 according to the second embodiment in that
Frame MV of color motion vector decoding device 2142
Output of converter 310b (frame-based motion vector)
Is not transmitted to the shape motion vector decoding device 2242.

The shape motion vector decoding device 2
In 242, when the predicted value of the shape motion vector of the macroblock to be processed in the shape image space is generated, if the reference macroblock in the color image space has been subjected to motion compensation processing in units of frames, Frame MV in motion vector decoding apparatus 2141
The corresponding motion vector stored in the memory 304a is referred to by the MV predictor 405b. On the other hand, when the reference macroblock in the color image space has been subjected to the motion compensation processing on a field basis, the MV predictor 4
In 05b, the output (motion vector in frame units) of the frame converter 310b of the color motion compensation encoding device 2142 is not referred to.

For this reason, the motion vector encoding apparatus 2042 according to the fourth embodiment does not refer to the color motion vector in the unit of a field, and is therefore different from the motion vector decoding apparatus 2041 according to the second embodiment (see FIG. 3). Therefore, there is a possibility that the prediction accuracy of the prediction value for the shape motion vector of the macroblock to be processed is deteriorated.

In the case where the macroblock to be processed has been subjected to motion compensation processing in frame units and the reference macroblock has been subjected to motion compensation processing in field units, color motion vector decoding is performed. Since the output of the frame MV converter 310b in the decoding device 2142 is referred to by the frame MV predictor 305a in the color motion vector decoding device 2142,
It can be said that the configuration of the motion vector encoding device according to the second embodiment is suitable.

That is, in the above case, the frame MV predictor 3 in the color motion vector decoding device 2142
Since it is necessary to refer to the motion vector in field units at 04a, the frame MV converter 310b performs a process of converting the motion vector in field units into frame units. Therefore, the frame MV converter 310b
Is output to the frame MV predictor 30 of the color motion vector decoding device 2142.
5a and the MV predictor 405b of the shape motion vector decoding device 2242, and the same is supplied to one of the frame MV predictors 305a, and the amount of processing in the frame MV converter 310b is the same. Therefore, it can be said that the configuration of the motion vector decoding apparatus according to the second embodiment is suitable for the above case from the viewpoint of the prediction accuracy of the predicted value of the shape motion vector.

However, for example, when both the macroblock to be processed and the reference macroblock in the color image space have been subjected to the motion compensation processing in units of fields, the color motion vector decoding device 2142
Then, the output of the field MV memory 304b is referred to. For this reason, in the configuration of the motion vector decoding device 2041 of the second embodiment that refers to the color motion vector of the reference macroblock on which the field-unit motion compensation processing has been performed in the prediction processing of the shape motion vector, the shape motion vector In order to perform the prediction process, the frame MV converter 310b must perform the process of converting the field-based color motion vector into the frame-based color motion vector.

On the other hand, in the motion vector decoding device 2142 of the fourth embodiment, only the output of the frame MV memory 304a of the shape motion vector coding device
05b.

With such a configuration, for example, if both the macroblock to be processed and the reference macroblock in the color image space have been subjected to motion compensation processing in units of fields, the shape motion vector encoding device 22
In 42, the motion vector of the reference macroblock is not referred to in the prediction processing of the shape motion vector, and as a result, the color motion vector coding device 2142
In the frame MV converter 310b, the process of converting a field-based color motion vector into a frame-based color motion vector can be omitted.

(Embodiment 5) FIG. 5 is a block diagram for explaining a motion vector encoding apparatus constituting an image encoding apparatus according to Embodiment 5 of the present invention. The motion vector encoding device 1073 according to the fifth embodiment includes a color motion vector encoding device 1173 that constitutes the color encoding device and a shape encoding device similar to the motion vector encoding device 1073 according to the third embodiment. And a shape motion vector encoding device 1273 that configures

Here, the color motion vector coding device 1173 has exactly the same configuration as the color motion vector coding device 1172 of the third embodiment. Further, the shape motion vector coding device 1273 is different from the color motion vector coding device 1127 in addition to the configuration of the shape motion vector coding device 1272 of the third embodiment.
72 frame MV memory 102a and MV predictor 204
b, the output of the frame MV memory 102a (that is, the motion vector in a frame unit) is supplied to the MV predictor 204b based on a non-interlace determination signal Nit from outside the image coding apparatus. It has a switch 220 for controlling. Therefore, the fifth embodiment
The configuration of the shape motion vector coding apparatus 1273 of the third embodiment is exactly the same as that of the shape motion vector coding apparatus 1272 of the third embodiment except for the switch 220. Here, the non-interlace determination signal Nit indicates whether the image signal input to the image encoding device is an interlace signal or a non-interlace signal.

Next, the function and effect will be described. When the image signal input to the image encoding device is non-interlaced, the switch 220 is turned on (conducting state) by the non-interlace determination signal Nit input to the input terminal 9a. Thereby, the frame MV memory 102 in the color motion vector encoding device 1173
The motion vector for each frame stored in a is stored in the MV predictor 20 in the shape motion vector encoder 1273.
4b.

Therefore, when a non-interlaced picture signal is input to the picture coding apparatus, the operation of the motion vector coding apparatus 1073 of the fifth embodiment is different from that of the conventional motion vector coding apparatus shown in FIG. It is completely the same, and the same coding efficiency can be realized.

On the other hand, when the image signal input to the image encoding device is an interlaced image signal, the switch 220 is turned on by the non-interlace determination signal Nit.
In the FF state (non-conduction state), the MV predictor 204b in the shape motion vector encoding device 1173 cannot refer to the color motion vector in the prediction processing.

For this reason, in the fifth embodiment, in the coding processing of the motion vector corresponding to the interlaced image signal, while the coding efficiency of the shape motion vector is deteriorated,
The MV predictor 204b does not need to determine whether a motion vector of a reference macroblock is a motion vector in a field unit or a motion vector in a frame unit as in the third embodiment. Can be reduced.

(Embodiment 6) FIG. 6 is a block diagram for explaining a motion vector decoding apparatus constituting an image decoding apparatus according to Embodiment 6 of the present invention. The motion vector decoding device 2043 according to the sixth embodiment corresponds to the motion vector coding device 1073 according to the fifth embodiment shown in FIG. This motion vector decoding device 204
3 is a color motion vector decoding device 2143 that decodes the color motion vector encoded signal Emvt output from the color motion vector encoding device 1173 in the motion vector encoding device 1073 and outputs the decoded signal Dmvt. , The motion vector encoding device 10
73, a shape motion vector decoding device 2243 that decodes the shape motion vector encoded signal Emvs output from the shape motion vector coding device 1273 and outputs the decoded signal Dmvs.

Here, the color motion vector decoding device 2143 has exactly the same configuration as the color motion vector decoding device 2142 of the fourth embodiment. Further, the shape motion vector decoding device 2243 includes the configuration of the shape motion vector decoding device 2242 of the fourth embodiment, and the color motion vector decoding device 2123.
73 frame MV memory 304a and MV predictor 405
b, the output of the frame MV memory 304a (that is, the motion vector for each frame) to the MV predictor 405b is supplied based on a non-interlace determination signal Nit from outside the image decoding apparatus. It is provided with a switch 400 for controlling. Therefore, Embodiment 6
The configuration of the shape motion vector decoding device 2243 of this embodiment is completely the same as that of the shape motion vector decoding device 2242 of the fourth embodiment except for the configuration of the switch 400. Here, the non-interlace determination signal Nit indicates whether or not the coded image signal input to the image decoding device is a coded signal of a non-interlaced image signal.

Next, the function and effect will be described. When an encoded signal of a non-interlaced image signal is input to the image decoding device, the switch 400 is turned on by the non-interlaced determination signal Nit input to the input terminal 9b.
State (conduction state). Thereby, the frame MV memory 3 in the color motion vector decoding device 2143
The motion vector for each frame stored in 04a can be referred to by the MV predictor 405b in the shape motion vector encoding device 2243.

Therefore, when a coded signal of a non-interlaced picture signal is input to the picture decoding apparatus, the operation of the motion vector coding apparatus 1073 of the sixth embodiment will be described with reference to FIG.
4 and is exactly the same as the conventional motion vector decoding device shown in FIG. 4, and the same decoding efficiency can be realized.

On the other hand, when the coded signal of the interlaced image signal is input to the image decoding device, the switch 400 is turned off by the non-interlace determination signal Nit.
The state becomes the F state (non-conduction state), and it becomes impossible for the MV predictor 405b in the shape motion vector decoding device 2143 to refer to the color motion vector in the prediction processing.

For this reason, in the sixth embodiment, in the decoding processing of the motion vector corresponding to the interlaced image signal, while the decoding efficiency of the shape motion vector is deteriorated,
The MV predictor 405b does not need to determine whether the motion vector of the reference macroblock is a motion vector in a field unit or a motion vector in a frame unit as in the fourth embodiment. Can be reduced.

An encoding or decoding program for realizing the configuration of the motion vector encoding device or the motion vector decoding device shown in each of the above embodiments is described below.
By recording the data on a storage medium such as a floppy disk, the processing described in each of the above embodiments can be easily performed by an independent computer system.

FIG. 7 shows the motion vector coding apparatus according to the first, third, or fifth embodiment or the motion vector decoding apparatus according to the second, fourth, or sixth embodiment. FIG. 11 is an explanatory diagram of a case where the present invention is implemented by a computer system using a disk.
Fig. 7 (a) shows the appearance of the floppy disk as viewed from the front,
FIG. 7 shows a sectional structure and a floppy disk body.
(b) shows an example of the physical format of the floppy disk body. The floppy disk FD has a structure in which the floppy disk main body D is accommodated in a floppy disk case FC. Tr is formed and each track T
r is divided into 16 sectors Se in the angular direction. Therefore, the floppy disk F storing the above program
In D, the floppy disk main body D has data as the program recorded in an area (sector) Se allocated thereon. FIG. 7C shows a configuration for recording the program on the floppy disk FD and performing image processing by software using the program stored in the floppy disk FD.

The above program is stored on a floppy disk FD
, The data as the program is written from the computer system Cs to the floppy disk FD via the floppy disk drive FDD.
When the image encoding device or the image decoding device is constructed in the computer system Cs by the program recorded on the floppy disk FD, the program is read from the floppy disk FD by the floppy disk drive FDD, and the computer system C
Load to s.

In the above description, the description has been made using the floppy disk as the data storage medium. However, even when the optical disk is used, the encoding process or the decoding process by the software may be performed in the same manner as in the case of the floppy disk. it can. Further, the data storage medium is not limited to the above-mentioned optical disk or floppy disk, but may be any type of IC card, ROM cassette, etc., as long as it can record a program. As in the case where a floppy disk or the like is used, encoding or decoding by software can be performed.

[0244]

As described above, the present invention (Claims 1, 7,
According to 13), color motion for converting a field-based color motion vector for performing a field-based motion compensation encoding process on an interlaced color signal for displaying an object in color into a frame-based color motion vector. In the encoding process for the shape motion vector for performing the frame-based motion compensation encoding process on the shape signal indicating the shape of the object including the vector conversion process, the prediction value of the shape motion vector corresponding to the macroblock to be processed is included. Is generated with reference to the shape motion vector and the color motion vector corresponding to the processed macroblock, and if the color motion vector corresponding to the processed macroblock is a field-based motion vector, the processed macroblock Color motion vector by the above color motion vector conversion process. Since the frame-based color motion vector obtained by the conversion is referred to, the coding process of the motion vector relating to the non-interlaced shape signal can be performed with the minimum circuit configuration extension. Satisfactorily in combination with the conversion treatment.

According to the present invention (claims 2, 8, and 14), in the encoding process for the shape motion vector, when the color motion vector corresponding to the processed macroblock is a frame-based motion vector, A predicted value of the shape motion vector corresponding to the block is generated with reference to the shape motion vector and the color motion vector corresponding to the processed macroblock, and the color motion vector corresponding to the processed macroblock is a field-based motion vector. At some time, since a predicted value of the shape motion vector corresponding to the processed macroblock is generated with reference to only the shape motion vector corresponding to the processed macroblock,
In the prediction process of the shape motion vector, the color motion vector in the field unit corresponding to the processed macro block is not referred to, and as a result, the color motion vector corresponding to the processed macro block and the processed macro block is not If it is a field-based motion vector,
In the encoding process for the color motion vector, the process of converting the field-based color motion vector into the frame-based color motion vector can be omitted.

According to the present invention (claims 3, 9, and 15), in the encoding process for the shape motion vector, when the input image signal is a non-interlaced image signal, the shape corresponding to the macroblock to be processed is determined. A prediction value of a motion vector is generated with reference to a shape motion vector and a color motion vector corresponding to a processed macroblock, and when the input image signal is an interlaced image signal, the prediction value corresponding to the processed macroblock is generated. Since the predicted value of the shape motion vector is generated with reference to only the shape motion vector corresponding to the processed macroblock, if the input image signal is an interlaced image signal, the coding process for the shape motion vector is performed. In, the color motion vector corresponding to the processed macroblock referred to is a field-based motion vector. Or determining a whether a motion vector of the frame is no longer necessary, it is possible to reduce the signal processing for this determination.

According to the present invention (claims 4, 10 and 16), it is possible to perform a field unit motion compensation decoding process on an encoded signal corresponding to an interlaced color signal for displaying an object in color. The method includes a color motion vector conversion process for converting a field-based color motion vector into a frame-based color motion vector, and in the decoding process on the shape motion vector, the prediction value of the shape motion vector corresponding to the macroblock to be processed is processed. When the color motion vector corresponding to the processed macroblock is generated by referring to the shape motion vector and the color motion vector corresponding to the processed macroblock, the color motion vector of the processed macroblock is used. Vector obtained by converting the vector by the above color motion vector conversion process. Since the frame-based color motion vector is referred to, the decoding process of the motion vector relating to the non-interlaced shape signal is preferably combined with the decoding process of the motion vector relating to the interlaced color signal with a minimum circuit configuration extension. Can be performed.

According to the present invention (claims 5, 11 and 17), in the decoding process for the shape motion vector, when the color motion vector corresponding to the processed macroblock is a frame-based motion vector, A predicted value of the shape motion vector corresponding to the block is generated with reference to the shape motion vector and the color motion vector corresponding to the processed macro block, and the color motion vector corresponding to the processed macro block is a field unit motion vector. , The predicted value of the shape motion vector corresponding to the macroblock to be processed is generated with reference to only the shape motion vector corresponding to the processed macroblock. The color motion vector in the field unit corresponding to the However, as a result, if the color motion vector corresponding to the processed macroblock and the processed macroblock is a field-based motion vector, the color motion vector decoding process converts the field-based color motion vector into a frame unit. Can be omitted.

According to the present invention (claims 6, 12 and 18), in the decoding processing for the shape motion vector, when the input image coded signal is an image coded signal corresponding to a non-interlaced image signal, Generating a predicted value of the shape motion vector corresponding to the processed macroblock with reference to the shape motion vector and the color motion vector corresponding to the processed macroblock, and the input image coded signal is an interlaced image signal. , The predicted value of the shape motion vector corresponding to the macroblock to be processed is generated with reference to only the shape motion vector corresponding to the processed macroblock. When the coded signal is an image coded signal corresponding to the interlaced image signal, in the decoding process for the shape motion vector, It is not necessary to motion vector corresponding to the processed macroblock irradiation to determine whether the motion vector in units of frames or a motion vector field units, it is possible to reduce the signal processing for this determination.

[Brief description of the drawings]

FIG. 1 is a block diagram for explaining an image encoding device according to Embodiment 1 of the present invention, and shows a motion vector encoding device constituting the image encoding device.

FIG. 2 is a block diagram for explaining an image decoding device according to Embodiment 2 of the present invention, and shows a motion vector decoding device that constitutes the image decoding device.

FIG. 3 is a block diagram for explaining an image encoding device according to Embodiment 3 of the present invention, and shows a motion vector encoding device constituting the image encoding device.

FIG. 4 is a block diagram for explaining an image decoding apparatus according to Embodiment 4 of the present invention, and shows a motion vector decoding apparatus constituting the image decoding apparatus.

FIG. 5 is a block diagram for explaining an image encoding device according to Embodiment 5 of the present invention, and shows a motion vector encoding device constituting the image encoding device.

FIG. 6 is a block diagram for explaining an image decoding apparatus according to Embodiment 6 of the present invention, and shows a motion vector decoding apparatus constituting the image decoding apparatus.

FIG. 7 is a data storage medium storing a program for performing a motion vector encoding process and a motion vector decoding process of each of the embodiments by a computer system (FIG.
(a), (b)) and the above computer system (Fig. (c)
FIG.

FIG. 8 is a schematic diagram for explaining an encoding process for each object based on MPEG4, and includes an image space Ts (FIG. (A)) obtained from a color signal and an image space Ss (FIG. b)), Color signal block processing (Fig.
(c), (e)), block processing of shape signal (Fig. (d),
(f)).

FIG. 9 is a block diagram for explaining a conventional image encoding apparatus conforming to MPEG4.

FIG. 10 is a block diagram for explaining a conventional image decoding device conforming to MPEG4.

FIG. 11 is a diagram for explaining the correspondence between macroblocks and motion vectors, and is a macroblock having four motion vectors (FIG. 11A) and a macroblock having two motion vectors (FIG. 11B). , A half macro block corresponding to each field in the macro block (FIG. (C)),
Macro block with one motion vector (Figure (d))
Is shown.

FIG. 12 is a diagram for explaining a conventional motion vector prediction process. In a case where a macroblock to be processed for a color signal has four motion vectors (FIG. 12A), the macroblock to be processed for a color signal is If the macroblock to be processed for the color signal has two motion vectors (FIG. 13C) when it has one motion vector (FIG. 13B), the motion vector prediction processing for the shape signal (FIG. 13D). Is shown.

FIG. 13 is a block diagram for explaining a conventional motion vector encoding device.

FIG. 14 is a block diagram for explaining a conventional motion vector decoding device.

FIG. 15 is a block diagram for explaining a conventional motion vector encoding device corresponding to an interlaced image signal.

FIG. 16 is a block diagram for explaining a conventional motion vector decoding device corresponding to an interlaced image signal.

[Explanation of symbols]

102a, 304a Frame MV memory 102b, 304b Field MV memory 103 MV effective memory 110a, 310a Field MV converter 110b, 310b Frame MV converter 104a, 305a Frame MV predictor 104b, 305b Field MV predictor 105a Frame MV encoder 105b Field MV encoder 303a Frame MV decoder 303b Field MV decoder 202, 404 MV memory 203, 402 MV effective memory 204a, 204b, 405a, 405b MV predictor 205 MV encoder 403 MV decoder 1000 Image coding device 1071, 1072, 1073 Motion vector coding device 1100 Color coding device 1171, 1172, 1173 Color motion vector code Apparatus 1200 Shape encoder 1271, 1272, 1273 Shape motion vector encoder 2000 Image decoder 2041, 2042, 2043 Motion vector decoder 2100 Color decoder 2141, 142, 2143 Color motion vector decoder 2200 Shape Decoding device 2241, 2242, 2243 Shape motion vector decoding device Cs Computer system D Floppy disk body FC Floppy disk case FD Floppy disk FD Floppy disk drive Se Sector Tr Track

Claims (18)

[Claims] An interlaced image signal corresponding to an individual object included in a predetermined image space and including a color signal for displaying the object in color and a shape signal indicating the shape of the object; An image coding apparatus that performs adaptive coding processing including frame-based and field-based motion compensation coding processing for each macroblock including a predetermined number of pixels that partition the image space. Encodes a color motion vector corresponding to a macroblock to be processed in a frame unit or a field unit for performing motion compensation coding processing in a frame unit or a field unit on a race color signal based on a predicted value thereof. And a frame-based motion compensation encoding process for an interlaced shape signal. And a shape motion vector encoding device that encodes a shape motion vector corresponding to a macroblock to be processed on the basis of a color motion vector corresponding to a processed macroblock and a predicted value obtained from the shape motion vector. The color motion vector encoding device has motion vector conversion means for converting a field-based color motion vector corresponding to a processed macroblock into a frame-based color motion vector, and the shape motion vector encoding device includes: When the color motion vector corresponding to the processed macroblock is a field-based motion vector, the predicted value of the shape motion vector corresponding to the processed macroblock is calculated by the frame-based shape corresponding to the processed macroblock. Motion vector and the above motion vector conversion means An image coding apparatus comprising: a shape motion vector predictor that generates a color motion vector for each frame output from the frame unit. 2. An image signal corresponding to an individual object included in a predetermined image space, the image signal including a color signal for displaying the object in color and a shape signal indicating a shape of the object. An image coding apparatus that performs adaptive coding processing including frame-based and field-based motion compensation coding processing for each macroblock including a predetermined number of pixels that partition the image space. Encodes a color motion vector corresponding to a macroblock to be processed in a frame unit or a field unit for performing a motion compensation coding process in a frame unit or a field unit for a race color signal based on a predicted value thereof. Color motion vector coding device and a frame-based motion compensation coding process for an interlaced shape signal. A shape motion vector encoding device that encodes a shape motion vector corresponding to a macroblock to be processed in frame units based on the prediction value thereof, wherein the shape motion vector encoding device comprises: When the color motion vector is a frame-based motion vector, a predicted value of the shape motion vector corresponding to the macroblock to be processed is generated based on the color motion vector and the shape motion vector corresponding to the processed macroblock, When the color motion vector of the processed macroblock is a field-based motion vector, the predicted value of the shape motion vector corresponding to the processed macroblock is changed to only the frame-based shape motion vector corresponding to the processed macroblock. Having a shape motion vector predictor that generates based on An image encoding device characterized by the following. 3. An interlaced or non-interlaced image signal corresponding to an individual object included in a predetermined image space and including a color signal for displaying the object in color and a shape signal indicating the shape of the object. An image encoding apparatus that performs adaptive encoding processing including motion compensation encoding processing in units of frames and fields on the image signal for each macroblock consisting of a predetermined number of pixels that divides the image space. There is a frame unit for performing a motion compensation coding process in a frame unit or a field unit for a color signal,
Alternatively, a color motion vector encoding device that encodes a color motion vector corresponding to a macroblock to be processed in a field unit based on a predicted value thereof, and a motion compensation encoding process in a frame unit for a non-interlaced shape signal. And a shape motion vector encoding device that encodes a shape motion vector corresponding to a macroblock to be processed on the basis of the prediction value in units of frames for performing When a non-interlaced image signal is received, the predicted value of the shape motion vector corresponding to the macroblock to be processed is
Generated based on the color motion vector and the shape motion vector corresponding to the processed macroblock, and when an interlaced image signal is received as the image signal, the predicted value of the shape motion vector corresponding to the processed macroblock is
An image coding apparatus comprising a shape motion vector predictor that generates a shape motion vector based only on a frame-based shape motion vector corresponding to a processed macroblock. 4. Image coding corresponding to an individual object included in a predetermined image space and corresponding to an interlaced image signal including a color signal for displaying the object in color and a shape signal indicating the shape of the object. Receiving the signal, and performing an adaptive decoding process including a motion compensation decoding process in a frame unit and a field unit on the image coded signal for each macroblock composed of a predetermined number of pixels for partitioning the image space. An image decoding apparatus, comprising: a frame unit or a field unit for performing a motion compensation decoding process in a frame unit or a field unit for an encoded signal corresponding to an interlaced color signal; A color motion vector decoding device for decoding a corresponding color motion vector based on the predicted value, and an interlaced shape signal The frame-based shape motion vector corresponding to the processed macroblock for performing the frame-based motion compensation decoding process on the encoded signal corresponding to the color signal and the shape motion vector corresponding to the processed macroblock. A shape motion vector decoding device for decoding based on a prediction value obtained from a vector, wherein the color motion vector decoding device converts a field-based color motion vector corresponding to the processed macroblock into a frame-based color motion vector. When the color motion vector corresponding to the processed macroblock is a field-based motion vector, the shape motion vector decoding device corresponds to the processed macroblock. The predicted value of the shape motion vector is Image decoding apparatus characterized by having a corresponding frame shape motion vector and a shape motion vector predictor which generates based on the color motion vector of a frame unit outputted from the motion vector conversion means. 5. Image coding corresponding to an individual object included in a predetermined image space and corresponding to an interlaced image signal including a color signal for displaying the object in color and a shape signal indicating the shape of the object. Receiving the signal, and performing an adaptive decoding process including a motion compensation decoding process in a frame unit and a field unit on the image coded signal for each macroblock composed of a predetermined number of pixels for partitioning the image space. An image decoding apparatus, comprising: a frame unit or a field unit for performing a motion compensation decoding process in a frame unit or a field unit for an encoded signal corresponding to an interlaced color signal; A color motion vector decoding device for decoding a corresponding color motion vector based on the predicted value, and an interlaced shape signal Shape motion vector decoding for decoding a frame-based shape motion vector corresponding to a macroblock to be processed for performing frame-based motion compensation decoding on an encoded signal corresponding to The shape motion vector decoding device, when the color motion vector of the processed macro block is a motion vector in units of frames, the shape motion vector prediction value corresponding to the processed macro block, Generated based on the color motion vector and the shape motion vector corresponding to the processed macroblock, and when the color motion vector of the processed macroblock is a field-based motion vector, the shape motion vector corresponding to the processed macroblock is generated. Of the predicted value of the frame unit corresponding to the processed macroblock An image decoding device comprising a shape motion vector predictor that generates a motion vector based only on a motion vector. 6. An interlaced or non-interlaced image signal corresponding to an individual object included in a predetermined image space and including a color signal for displaying the object in color and a shape signal indicating the shape of the object. And performs adaptive decoding processing including frame-based and field-based motion-compensated decoding processing on the image-encoded signal to a predetermined number of pixels for partitioning the image space. An image decoding apparatus for each macro block, comprising: a macro block to be processed in a frame unit or a field unit for performing a motion compensation decoding process in a frame unit or a field unit on an encoded signal of a color signal. A color motion vector decoding device that encodes a color motion vector corresponding to Motion vector decoding for decoding, based on the predicted value, a shape motion vector corresponding to a macroblock to be processed in a frame unit for performing a motion compensation decoding process in a frame unit for a coded signal of a signal. The shape motion vector decoding device predicts a shape motion vector corresponding to the macroblock to be processed when receiving an image coded signal corresponding to a non-interlaced image signal as the image coded signal. Generating a value based on the color motion vector and the shape motion vector corresponding to the processed macroblock, and receiving the image coded signal corresponding to the interlaced image signal as the image coded signal. The predicted value of the shape motion vector corresponding to the Image decoding apparatus characterized by having a shape motion vector predictor which generates based on Runomi. 7. The image space is divided by an adaptive coding process including a motion compensation coding process in a frame unit and a field unit for interlaced image signals corresponding to individual objects included in a predetermined image space. An image coding method for coding for each macroblock consisting of a predetermined number of pixels to perform a motion compensation coding process for each frame on an interlaced shape signal indicating a shape of an object included in the image signal. Motion vector coding processing for coding a macro motion vector for each macroblock based on the predicted value, and a field for an interlaced color signal for displaying an object included in the image signal in color. Convert field-based color motion vectors to frame-based color motion vectors for unit-based motion compensation coding In the shape motion vector encoding process, the predicted value of the shape motion vector corresponding to the macroblock to be processed is referred to the shape motion vector and the color motion vector corresponding to the processed macroblock. When the color motion vector corresponding to the processed macroblock is a field-based motion vector, the color motion vector of the processed macroblock is converted by the color motion vector conversion process. An image encoding method characterized by referring to a color motion vector in a frame unit. 8. The image space is divided by an adaptive encoding process including a frame-based and a field-based motion compensation encoding process for interlaced image signals corresponding to individual objects included in a predetermined image space. An image coding method for coding for each macroblock consisting of a predetermined number of pixels, wherein the motion of the interlaced color signal for color-displaying the object included in the image signal is performed in frame units or field units. A color motion vector encoding process of encoding a color motion vector in a frame unit or a field unit for performing a compensation encoding process for each macroblock based on a prediction value thereof; A shape motion vector for performing frame-based motion compensation coding on an interlaced shape signal indicating the shape. And a shape motion vector coding process for coding each of the macro blocks based on the predicted value for each macroblock. In the shape motion vector coding process, the color motion vector corresponding to the processed macroblock is frame-by-frame. , The predicted value of the shape motion vector corresponding to the macroblock to be processed is
When the color motion vector corresponding to the processed macroblock is generated by referring to the shape motion vector and the color motion vector corresponding to the processed macroblock, and the color motion vector corresponding to the processed macroblock is a field-based motion vector, the shape corresponding to the processed macroblock is generated. An image coding method, wherein a predicted value of a motion vector is generated with reference to only a shape motion vector corresponding to the processed macroblock. 9. An interlaced or non-interlaced image signal corresponding to an individual object included in a predetermined image space is subjected to adaptive encoding processing including frame-based and field-based motion compensation encoding processing. What is claimed is: 1. An image coding method for coding a macroblock consisting of a predetermined number of pixels for partitioning an image space, comprising: a frame signal or a field unit for a color signal for displaying an object included in the image signal in color. A color motion vector encoding process for encoding a frame unit or a field unit color motion vector for each macroblock based on the predicted value thereof, for performing the motion compensation encoding process of To perform frame-based motion compensation coding on non-interlaced shape signals indicating the shape of the object A shape motion vector encoding process for encoding a shape motion vector for each macroblock based on the predicted value thereof. In the shape motion vector encoding process, when the image signal is a non-interlaced image signal, Generating a predicted value of the shape motion vector corresponding to the processed macroblock with reference to the shape motion vector and the color motion vector corresponding to the processed macroblock, and when the image signal is an interlaced image signal, An image coding method, wherein a predicted value of a shape motion vector corresponding to a processed macroblock is generated with reference to only a shape motion vector corresponding to a processed macroblock. 10. An adaptive decoding process including a motion compensation decoding process in a frame unit and a field unit for an image coded signal corresponding to an interlaced image signal corresponding to an individual object included in a predetermined image space. An image decoding method for decoding for each macroblock consisting of a predetermined number of pixels for partitioning the image space, wherein an encoded signal corresponding to an interlaced shape signal indicating a shape of an object included in the image signal A shape motion vector for performing frame-based motion compensation decoding on
A shape motion vector decoding process for decoding each macroblock based on the predicted value, and a field unit for a coded signal corresponding to an interlaced color signal for displaying an object included in the image signal in color. And a color motion vector conversion process of converting a field-based color motion vector into a frame-based color motion vector for performing the motion compensation decoding process of the above. When the predicted value of the shape motion vector to be generated is generated with reference to the shape motion vector and the color motion vector corresponding to the processed macro block, the color motion vector corresponding to the processed macro block is a field unit motion vector. In this case, the color motion vector of the processed macroblock is Image decoding method characterized by reference to the color motion vector of each frame obtained by converting a color motion vector conversion process. 11. An image encoding signal for an interlaced image signal corresponding to an individual object included in a predetermined image space is adaptively decoded by a frame-based and a field-based motion-compensated decoding process. What is claimed is: 1. An image decoding method for decoding a macroblock consisting of a predetermined number of pixels for partitioning an image space, wherein the encoded signal corresponds to an interlaced color signal for color-displaying an object included in the image signal. Color motion vector decoding for decoding a frame-based or field-based color motion vector for performing frame-based or field-based motion compensation decoding processing for each macroblock based on its predicted value Processing, and an encoded signal corresponding to an interlaced shape signal indicating the shape of an object included in the image signal. The shape motion vector for performing motion compensation decoding process in units of frames with respect to,
And a shape motion vector decoding process for decoding each macroblock based on the predicted value. In the shape motion vector decoding process, the color motion vector corresponding to the processed macroblock is a frame-based motion vector. , The predicted value of the shape motion vector corresponding to the macroblock to be processed is
When the color motion vector corresponding to the processed macroblock is generated by referring to the shape motion vector and the color motion vector corresponding to the processed macroblock, and the color motion vector corresponding to the processed macroblock is a field-based motion vector, the shape corresponding to the processed macroblock is generated. An image decoding method, wherein a predicted value of a motion vector is generated with reference to only a shape motion vector corresponding to the processed macroblock. 12. An adaptive decoding including a frame-based and field-based motion-compensated decoding process for an image-encoded signal corresponding to an interlaced or non-interlaced image signal corresponding to an individual object included in a predetermined image space. An image decoding method for decoding, for each macroblock consisting of a predetermined number of pixels that divides the image space, into a code corresponding to a color signal for color-displaying an object included in the image signal. Frame unit for the converted signal,
Alternatively, a color motion vector decoding process for decoding a frame unit or a field unit color motion vector for performing a field unit motion compensation decoding process for each macroblock based on a predicted value thereof; A shape motion vector for performing a frame-based motion compensation decoding process on a coded signal corresponding to a non-interlaced shape signal indicating the shape of an object included in each macroblock based on its predicted value. And a shape motion vector decoding process for decoding. In the shape motion vector decoding process, when the image coded signal is an image coded signal corresponding to a non-interlaced image signal, The predicted value of the shape motion vector to be processed is calculated based on the shape motion vector corresponding to the processed macroblock and the color motion vector. When the image coded signal is an image coded signal corresponding to an interlaced image signal, the predicted value of the shape motion vector corresponding to the processed macroblock is generated by referring to the processed macroblock. An image decoding method characterized in that the image decoding method is generated by referring to only a shape motion vector to be processed. 13. A data storage medium storing an image processing program, wherein the image processing program is an encoding program for causing a computer to perform an encoding process according to the image encoding method according to claim 7. A data storage medium characterized by the following. 14. A data storage medium storing an image processing program, wherein the image processing program is an encoding program for causing a computer to perform an encoding process according to the image encoding method according to claim 8. A data storage medium characterized by the following. 15. A data storage medium storing an image processing program, wherein the image processing program is an encoding program for causing a computer to perform an encoding process according to the image encoding method according to claim 9. A data storage medium characterized by the following. 16. A data storage medium storing an image processing program, wherein the image processing program is a decoding program for causing a computer to perform a decoding process according to the image decoding method according to claim 10. A data storage medium characterized by the following. 17. A data storage medium storing an image processing program, wherein the image processing program is a decoding program for causing a computer to perform a decoding process according to the image decoding method according to claim 11. A data storage medium characterized by the following. 18. A data storage medium storing an image processing program, wherein the image processing program is a decoding program for causing a computer to perform a decoding process according to the image decoding method according to claim 12. A data storage medium characterized by the following.