KR100275270B1 - Method and apparatus for interlaced binary shape coding - Google Patents

Abstract

PURPOSE: An apparatus is provided to be capable of decoding a binary shape signal. CONSTITUTION: A field binary alpha block(BAB) mode sensing part(10) checks whether a coding mode of a current field BAB is "all_0" or "all_255. If the coding mode is neither "all_0" nor "all_255", the field BAB mode sensing part(10) supplies the current field BAB data to a field BAB mode select part(20), a motion estimation part(32) in a motion estimating and compensating part(30), the first multiplexer(72) in a field reproduction part(70), via a path(L10). The field BAB mode select part(20) is supplied with a motion-compensated field BAB data from a motion compensation part(34) in the motion estimating and compensating part(30) via a path(L30). A field memory(74) stores reproduced field data, which indicates information associated with two field coded just before a current field. If the current field BAB is an upper field BAB included in a current upper field, the motion estimation part(32) extracts previously reproduced lower field data and previously reproduced upper field data from the field memory(74) via a path(L70). If the current field BAB is a lower field BAB included in a current lower field, the motion estimation part(32) extracts previously reproduced lower field data and presently reproduced upper field data from the field memory(74) via a path(L70). The current field BAB data is applied to the motion estimation part(32) from the field BAB mode sensing part(10) via the path(L10).

Description

A method and apparatus for joint encoding of binary shape signals

The present invention relates to a method and apparatus for encoding a binary shape signal, and more particularly, to a method and apparatus for encoding a binary shape signal.

Binary shape signals representing the position and shape of an object can be represented as a binary alpha block (BAB) within a frame (or VOP), for example, a 16 × 16 binary pixel size, each binary pixel being, for example, zero. Or a binary value representing a background pixel or an object pixel, such as 255. BAB may be encoded using a shape coding technique based on a general bitmap such as a context-based arithmetic encoding (CAE) technique.

For example, in an intra-frame, the current BAB is encoded using the general intra-CAE technique, where each pixel in the current BAB is arithmetically based on an intra-context composed of a set of pixels selected from the current frame. Is encoded. In an inter-frame, the current BAB is encoded using intra-CAE or inter-CAE techniques depending on which CAE technique produces less amount of encoded data. According to the inter-CAE technique, an error representing the difference between the current BAB and each of the predetermined candidate BABs included in the previous frame is first calculated, and the most similar candidate BABa and the motion vector are obtained by the motion estimation technique. A similar candidate BAB represents a candidate BAB that produces the least error among the candidate BABs, and the motion vector represents a displacement between the current BAB and the most similar candidate BAB. Accordingly, each pixel of the current BAB is encoded using, for example, a variable length coding (VLC) technique based on the inter-context and the motion vector difference indicating the difference between the motion vector and the motion vector estimate. The inter-context consists of a set of pixels selected in the current frame and a set of pixels selected from the previous frame based on the motion vector. For CAE techniques and motion vectors, see MPEG-4 Video Verificatioh Model Version 7.0, International Organization for Standardization, Coding of Moving Picture and Associated Audio Information, ISO / IEC JTC1 / SC29 / WG11 MPEG97 / N1642, Bristol, April 1997.

In order to encode binary shape signals more efficiently, one of the mode signals in Table 1 is assigned to each BAB.

Mode 1 MVD == 0 & & No Update Mode 2 MVD! = 0 ＆＆ No UPdate Mode 3 ALL 0 Mode 4 ALL 255 Mode 5 intra-CAE Mode 6 MVD == 0 & & inter-CAE Mode 7 MVD! = 0 & & inter-CAE

Mode 1 means that the Motion Vector Difference (MVD) of BAB is 0 and BAB can be represented by the closest candidate BAB, Mode 2 means that the MVD of BAB is non-zero and BAB is the closest candidate BAB It can be represented. Only the mode signal is encoded for BAB in mode 1, and BAB in mode 2 is represented by the mode signal and MVD. In determining " No Update ", a difference BAB consisting of the difference between each pixel of the most similar BAB and the corresponding pixel currently in the BAB is formed, and the error of the sub block is, for example, the absolute value of the pixels in the sub block. When summed, it is checked whether an error of 4x4 subblocks of 4x4 pixel size included in the difference BAB is smaller than a predetermined reference value. If the errors of all sub blocks are less than or equal to the reference value, the BAB is determined to be Mode 1 or Mode 2 depending on the MVD value.

Similarly, if the error for the 4x4 subblock is all less than or equal to the reference value when the pixels in the BAB all change to zero, then the BAB is encoded in all_0 mode, ie mode 3. If the errors in the 4x4 subblock are all less than or equal to the reference value when the pixels in the BAB all change to 255, the BAB is encoded in all_255 mode, that is, mode 4. For the BAB in mode 3 or 4, only the mode signal needs to be encoded. If the BAB does not fall between modes 1 to 4, "intra-CAE" or "inter-CAE" is employed to encode the BAB, and the mode 5 BAB is represented by the mode signal and the intra-CAE coded BAB data. BAB in mode 6 is represented by the mode signal and the inter-CAE coded BAB data; BAB in mode 7 is represented by a mode signal, inter-CAE coded BAB data, MVD.

In MPEG-4, the mode determination method described above has been proposed for encoding a binary shape signal by frame rather than field-by-field encoding, and a method and apparatus for encoding a binary shape signal by field have not been invented. In encoding the binary shape signal frame by frame, if the motion of an object in one frame or VOP is large, the coding efficiency is reduced. Referring to Figure 1a to 1c, there is shown an object having a large movement, which is represented by a field and a frame, respectively.

1a shows the upper field roll; 1B shows the bottom field; Figure 1C shows a frame of binary shape signals, where each square in the field and frame represents one pixel, the hatched squares are pixels belonging to the object, and the white squares are pixels belonging to the background. The frame is obtained by alternating each row of fields, that is, for example, the rows of the upper field are located in the even rows of the frame with rows 0 through (2N-1), and the rows of the lower fields are framed. Located in odd rows of, each field has N rows that are positive integers. The general binary shape signal coding method employing the progressive coding method encodes a binary shape signal frame by frame. However, when the motion of an object in a frame is large as shown in Figs. 1A to 1C, a binary shape signal having very little spatial correlation is encoded, resulting in a decrease in coding efficiency.

So far, no prior art has invented a method and apparatus capable of encoding a binary shape signal field by field.

It is therefore an object of the present invention to provide a method and apparatus for encoding a binary shape signal.

Another object of the present invention is to improve a coding efficiency by providing a method and apparatus for encoding a binary shape signal for each field.

In order to achieve the above object, according to an aspect of the present invention, in a method for encoding a target block of a binary shape signal, the binary shape signal includes a plurality of pictures, each picture being a first and a second binary number. When a target block is one of the blocks of the current picture to be encoded, and M and N are positive integers: (a) First and second reference blocks If the error of the target block for is not greater than a predetermined reference value, the mode of the target block is determined as the first and second modes, respectively, wherein each reference block has M × N pixels, and the first and second references are referred to. All pixels of the block have first and second binary values, respectively; (b) Motion vector information and motion compensation including a motion vector by motion estimation and compensation of the target block for one or more previous pictures for the current picture, if not determined as the first and second modes in step (a). Generating a block, wherein the motion compensation block includes a block most similar to the target block; (c) calculating a motion compensation error (MCE) between the target block and the most similar block and a motion vector difference (MVD) between the motion vector and its estimate; (d) if the MVD is not 0, encoding the MVD to provide encoded MVD data; (e) if the MVD is 0 and MCE is not greater than the reference value, set the mode of the target block to the third mode; if the MVD has a nonzero value and MCE is not greater than the reference value, set the mode of the target block to the fourth mode; ; (f) if not determined as the third or fourth mode in step (e), generate intra-encoded data and inter-encoded data, the intra-encoded data being based on the pixels of the predetermined current picture; Generating pixels by encoding the pixels of the target block, and encoding the pixels of the target block based on predetermined pixels included in the current picture and the motion compensation block; (g) comparing the number of bits of intra-coded data with the number of bits of inter-coded data; (h) if the number of bits of the intra-coded data is not greater than the number of bits of the inter-coded data, setting the mode of the target block to the fifth mode; (i) if it is not determined as the fifth mode in step (h), if the MVD is 0, setting the mode of the target block to the sixth mode; and if the MVD is not 0, setting the mode of the target block to the seventh mode. A binary shape signal encoding method is provided.

In order to achieve the above object, according to another aspect of the present invention, in an apparatus for encoding a target block of a binary shape signal, the binary shape signal includes a plurality of pictures, each of the pictures are first and second When the target block is one of the blocks of the current picture to be encoded, and M and N are positive integers: see first and second of the target block. If the error for the block is not greater than a predetermined reference value, a first identification signal of type 1 and type 2 is generated for the target block, respectively, wherein each reference block has M × N pixels, Mode sensing means having all pixels of the second reference block having first and second binary values, respectively; Reproducing means for reproducing the encoded binary shape signal and for generating and storing the reproduced binary shape signal, wherein the reproduced binary shape signal includes previously coded pictures; The motion vector information and the motion compensation block are generated by motion estimation of the target block on the reproduced binary shape signal corresponding to one or more previous pictures for the current picture, wherein the motion vector information includes the motion vector, and the motion compensation block is Motion estimation and compensation means including a block most similar to the target block; Mode selection means for calculating a motion compensation error (MCE) between the target block and the closest block and generating a second identification signal if the MCE is not greater than a reference value; Motion vector difference (MVD) calculating means for calculating a motion vector difference (MVD) between the motion vector and its estimate and encoding MVD to generate coded MVD data; Intra- and inter-encoding data is performed on the target block based on the target block, the motion compensation block, and the reproduced binary shape signal to generate intra- and inter-coded data, respectively, and the number of bits of the intra- and inter-coded data. Select one of the intra- and inter-encoded data based on the second reference signal; provide a type 3 third identification signal if the intra-encoded data is selected, and provide a type 4 third identification signal if the inter-encoded data is selected Encoding means for performing; According to the MVD, the encoded MVD, the encoded data and the identification signals, the binary shape signal encoding apparatus comprising a formatting means for determining the target block mode and generating the encoded target block data according to the determined mode do.

1a, b and c show an upper field, a lower field and a frame in a binary shape signal, respectively;

2 is a block diagram showing the apparatus for encoding a binary form signal according to the present invention;

3 is a diagram illustrating a motion estimation process of a field binary alpha block (BAB) according to the present invention;

4 is a flowchart illustrating an encoding process according to the present invention.

<Description of the code | symbol about the principal part of drawing>

10: field BAB mode detection unit 20: field BAB mode selection unit

30: motion estimation and compensation unit 32: motion estimation unit

34: motion compensation unit 40: motion vector (MV) determination unit

42: MV memory 44: Motion vector difference (MVD) calculation unit

46: MVD encoder 50: context-based arithmetic coding (CAE) unit

52-1: Intra CAE unit 54-1: Intra bit calculator

52-2: Inter CAE unit 54-2: Inter bit calculation unit

56: comparison unit 58: selection unit

60: formatting section 62: mode determination section

64 mode encoding unit 66 second multiplexing unit

70: field reproducing unit 72: first multiplexer

74: field memory

According to the present invention, there is provided a method and apparatus for effectively encoding a binary shape signal for each picture, wherein a picture represents a frame or a field. According to a preferred embodiment of the present invention, the picture is treated as a field, and the present invention will be described for encoding binary shape signals on a field-by-field basis to accommodate large movements of an object. The binary shape signal includes the previous and current frames, the previous frame is divided into the previous upper and previous lower fields, and the current frame is divided into the current upper and current lower fields. The method and apparatus according to the present invention processes the fields in the order of the previous top field, the previous bottom field, the current top field, and the current bottom field. Each field is divided into blocks of P × Q binary pixels when P and Q are positive integers, each block referred to as an upper field BAB or a lower field BAB. In a preferred embodiment of the present invention, both P and Q are set to 16, and in another embodiment, P and Q are set to 8 and 16.

Referring to Fig. 2, there is shown an apparatus for encoding the binary shape signal in accordance with the present invention. Current upper field BAB data or current lower BAB data is input to the field BAB mode sensing unit 10, wherein the current upper field data includes binary pixel data of the current upper field BAB; The current lower field BAB contains binary pixel data of the current lower field BAB. In binary appeal data, objects and background pixels are represented by binary numbers, for example, 255 and 0, respectively.

The field BAB mode detection unit 10 checks whether the encoding mode of the current field BAB is "all_0" or "all_255". Specifically, the current field BAB is divided into a T x S pixel size, for example, a 4 x 4 pixel size when T and S are positive integers. A field BAB having a size of 16 x 16 pixels in the first embodiment is 4 x 4 A field BAB including 8 subblocks and 8 × 16 pixels in size includes 2 × 4 subblocks. If the error between the subblock of the current field BAB and the subblock of the all_0 field BAB is less than or equal to a predetermined reference value, an identification signal S1 of type 1 indicating that the encoding mode of the current field BAB is "all_0" is generated and the formatting unit 60 is performed. ) Is provided to the mode determining unit 62, in which the all_0 field BAB is a field BAB in which all pixel values are zero. If the error between the subblock of the current field BAB and the subblock of the all_255 field BAB is less than or equal to a predetermined reference value, an identification signal S1 of type 2 indicating that the encoding mode of the current field BAB is "all_255" is generated and the formatting unit 60 is used. Is supplied to the mode determining unit 62, in which the all_255 field BAB is a field BAB in which all pixel values are 255.

If the encoding mode of the current field BAB is neither "all_0" nor "all_255", the field BAB mode detection unit 10 transmits the current field BAB data into the field BAB mode selection unit 20, the motion estimation and compensation unit 30. To the first multiplexer 72 in the motion estimation section 32 and the field reproducing section 70 via the path L10. The field BAB mode selector 20 receives the motion compensated field BAB data from the motion compensator 34 in the motion estimation and compensator 30 through the path L30. The motion estimation and compensation process will be described in two cases, where the current field BAB is the upper field BAB and the lower field BAB.

In the field memory 74 in the field reproducing section 70, reproduced field data is stored, and the reproduced field data indicates information on two fields encoded immediately before the current field. If the current field BAB is the upper field BAB included in the current upper field, the motion estimation unit 32 extracts from the field memory 74 the previous upper field data and the previous lower field data reproduced through the path L70; If the current field BAB is the lower field BAB included in the current lower field, the motion estimation unit 32 extracts the previous lower field data reproduced through the path L70 and the reproduced current upper field data from the field memory 74. The current field BAB data is applied from the field BAB mode detector 10 to the motion estimation unit 32 via the path L10.

Referring to FIG. 3, a motion estimation and compensation process according to a preferred embodiment of the present invention is shown. The current frame 100 includes the current upper field 110 and the current lower field 120, and the previous frame 200 includes the previous upper field 210 and the previous lower field 220.

If the current field BAB is the upper field BAB included in the current upper field 110, for example, the current upper field BAB 112, then the motion estimator 32 is the first previous lower field BAB 222 within the previous lower field 220. The first previous lower field BAB 222 is in the same position as the current upper field BAB 112 in the current upper field 110 in the previous lower field 220. Then, the current upper field BAB 112 is moved by one pixel within the first previous lower search area 226 formed in the previous lower field 220, where the first previous lower search area 226 is moved to the first previous lower area. The field 222 includes a plurality of candidate field BABs. At each displacement, the error between the current upper field BAB 112 and the corresponding candidate field BAB is calculated.

The motion estimator 32 then detects the previous upper field BAB 212 in the previous upper field 210, where the previous upper field BAB 212 is at the same position as the current upper field BAB 112. Then, the current upper field BAB 112 is moved by one pixel in the previous upper search area 214 formed in the previous upper field 210, where the previous upper search area 214 moves the previous upper field BAB 212. Including a plurality of candidate field BABs. The error between the current upper field BAB 112 and the corresponding candidate field BAB at each displacement is calculated.

Similarly, if the current field BAB is a lower field BAB included in the current lower field 120, for example, the current lower field BAB 122, the motion estimation unit 32 may determine the current upper field BAB (which is in the current upper field 110). 114, the current upper field BAB 114 is now in the same position as the lower field BAB 122. Then, the current lower field BAB 122 is moved by one pixel within the current upper navigation area 116 formed in the current upper field 110, where the current upper navigation area 116 includes the current upper field 114. Thus including a number of candidate field BABs. At each displacement, the error between the current lower field BAB 122 and the corresponding candidate field BAB is calculated.

The motion estimator 32 then detects the previous lower field BAB 224 within the previous lower field 220, where the previous lower field BAB 224 is at the same position as the current lower field BAB 122. Then, the current lower field BAB 122 is moved by one pixel in the second previous lower search area 228 formed in the previous lower field 220, where the second previous lower search area 2228 is moved to the previous lower field BAB. A plurality of candidate field BABs are included, including 224. The error between the current lower field BAB 122 and the corresponding candidate field BAB at each displacement is calculated.

As depicted above, the motion estimator 32 selects as the best candidate field BAB the candidate field BAB that motion estimates the current field BAB for the previous two fields and generates the minimum error. The output of the motion estimation unit 32 calculates the motion vector difference (MVD) and the motion vector (MV) memory 42 in the motion vector difference (MVD) determination unit 40 as the current motion vector and the field identification flag of the path L34. Applied to the unit 44 and the motion estimation unit 34, the current motion vector indicates a displacement between the current field BAB and the best candidate field BAB, and the field identification flag indicates which field the best candidate field BAB belongs to.

The motion compensator 34 extracts the framed BAB data framed from the field memory 74 via the path L72, and the framed field BAB data is framed at the optimal candidate field BAB and the width of one pixel around the frame. To say; The framed field BAB data is provided through the path L30 to the field BAB mode selector 20 and the first multiplexer 72 as motion compensated field BAB data.

The MVD calculator 44 extracts a motion vector estimate from the MV memory 42 according to the current motion vector and the field identification flag input from the motion estimator 32 via the path L34. The motion vector of one of the predetermined neighboring field BABs of the current field BAB determined according to four. The motion vector difference (MVD) between the current motion vector and the corresponding motion vector estimate is then calculated. The MVD and field identification flags are provided to the MVD encoder 46 and the mode determiner 62 in the motion vector difference determiner 40.

The MVD encoder 46 encodes the field identification flags from the MVD and the MVD calculator 44 only when the MVD has a non-zero value, and if the encoded MVD exists, the encoded MVD and the encoded field identification flag. To the second multiplexer 66 in the inter-bit computation unit 54-2 in the context-based arithmetic encoding (CAE) unit 50 and the selector 58 and the formatting unit 60 as encoded MVD data. It is provided through the path L40, CAE unit 50 is the intra-CAE unit 52-1 and the inter-CAE unit 52-2, the intra-bit calculation unit 54-1 and the inter-bit calculation unit ( 54-2), a comparator 56 and a selector 58.

On the other hand, according to the current field BAB data of the path L10 and the motion compensated field BAB of the path L30, the field BAB mode selector 20 selects 4 × 4 an optimal candidate field BAB included in the current field BAB and the motion compensated field BAB. Divide into subblocks of pixel size. If the error between the subblock in the current field BAB and the corresponding candidate block in the best candidate field BAB is less than or equal to the predetermined reference value, the field BAB mode selector 20 generates an identification signal S2, which is the current BAB field. Denotes that there is no need to be coded, and S2 is provided to the mode determination unit 62.

If the error between the current field BAB and the corresponding sub-block in the best candidate field BAB is greater than a predetermined reference value, the field BAB mode selection unit 20 intercepts the current field BAB data with the intra-CAE unit 52-1. The CAE unit 52-2 provides the motion compensated field BAB data to the inter-CAE unit 52-2.

The intra-CAE unit 52-1 is in the current field BAB based on the current field BAB itself and previously reproduced field BABs located around the current field BAB input via the path L76 from the first multiplexer 72. Generate an intra-context for each pixel; Based on the generated intra-context, the current field BAB is encoded using a general intra-CAE technique. The intra-CAE for the current field BAB is input to the intra-bit calculator 54-1 and the selector 58.

The intra-bit calculator 54-1 calculates the number of bits of the intra-CAE data and provides the calculated number of bits to the comparator 56. FIG.

The inter-CAE section 52-2 intersects each pixel in the current field BAB based on the previously reproduced field BABs on the path L76 and the motion compensated field BAB of the field BAB mode selector 20. Create a context; The current field BAB is encoded based on the general inter-CAE technique. The inter-CAE data for the current field BAB is applied to the inter-bit calculator 54-2 and the selector 58.

The interbit calculation unit 54-2 calculates the number of bits of the encoded MVD data and the inter-CAE data based on the encoded MVD data on the path L40 and the inter-CAE data from the inter-CAE unit 52-2. The calculated and calculated number of bits is applied to the comparator 56.

The comparison unit 560 compares the number of bits of the intra-CAE data with the number of bits of the inter-CAE data and the encoded MVD data. The number of bits of the intra-CAE data is smaller than the number of bits of the inter-CAE data and the encoded MVD data. Or the same, the comparator 56 provides the type 3 identification signal S3 to the selection unit 58 and the mode determination unit 62; otherwise, the type 4 identification signal S3 is selected for the selection unit 58 and the mode determination unit. Provided at 62.

The selector 58 selects the intra-CAE data or the inter-CAE data and the encoded MVD data according to the type 3 or 4 identification signal and provides the same to the second multiplexer 66 in the formatting unit 60.

The mode determiner 62 is provided with identification signals S1, S2, S3 and MVD input from the field BAB mode detector 10, the field BAB mode selector 20, the comparator 56 and the MVD calculator 44, respectively. The coding mode of the current field BAB is determined based on the method, and the determined mode signal is provided to the mode encoder 64, the second multiplexer 66, and the first multiplexer 72 in the formatting unit 60. . Table 2 shows a method of determining an encoding mode of a current field BAB according to the present invention.

In particular, mode 1 is determined for the current field BAB when S2 is present and MVD is zero; Mode 2 is when S2 is present and MVD is not zero; Mode 3 is when type 1 S1 is present; Mode 4 is when S1 of type 2 is present; Mode 5 is when type 3 S3 is present regardless of the MVD value; Mode 6 is when type 4 S3 is present and MVD is 0; Mode 7 is when S3 of type 4 is present and MVD is not zero, and the "x" mark indicates that the corresponding signal is ineffective.

S1 S2 S3 MVD mode × ○ × = 0 Mode 1 × ○ × ≠ 0 Mode 2 Type 1 × × × Mode 3 Type 2 × × × Mode 4 × × Type 3 = 0 or ≠ 0 Mode 5 × × Type 4 = 0 Mode 6 × × Type 4 ≠ 0 Mode 7

The first multiplexer 72 reproduces the field BAB data corresponding to the current field BAB in accordance with the mode signal from the mode determiner 62. In other words, the first multiplexer 72 provides the all_0 field BAB to the field memory 74 via the path L76 according to the mode 3 signal; All_255 field BAB according to the mode 4 signal; Current field BAB according to mode 5, 6, or 7 signal; The best candidate field BAB is provided according to the mode 1 or 2 signal.

The mode encoder 64 encodes the mode signal from the mode determiner 62 using a general VLC technique, and provides the encoded mode signal to the second multiplexer 66.

In accordance with the mode signal from the mode determiner 66, the second multiplexer 66 selectively multiplexes the input signal and transmits the multiplexed signal to the transmitter (not shown) as encoded current field BAB data. to provide. The coded current field BAB is a coded mode signal and coded MVD data when it is a mode 1 or 2 signal; In the case of a mode 3 or 4 signal, it is an encoded mode signal; In the case of a mode 5 signal, it is intra-CAE data; In the case of a mode 6 or 7 signal, it is a coded mode signal, inter-CAE data, and coded MVD data. Note that when MVD is 0 as in the case of Mode 1 or Mode 6, the encoded MVD data includes the encoded field identification signal.

4A and 4B, a process of encoding a binary shape signal for each field according to the present invention is illustrated. In step S1, it is checked whether the pixels of the current field BAB can be replaced with the mode 255 or 0, and in step S2, as in the field BAB mode sensing unit 10 shown in FIG. Mode 3 if all can be replaced by 0, mode 4 determines the mode of the current field BAB if all pixels of the current field BAB can be replaced with 255. If the mode of the current field BAB is determined to be mode 3 or 4, in step S3 the mode signal is encoded and ends by transmitting the encoded mode signal.

If the result of the investigation in step S1 is negative, go to step S4, the motion estimation and compensation for the current field BAB is performed in the motion estimation and compensation unit 30 shown in FIG. And a field identification flag (FIF). In step S5, the motion compensation error MCE and MVD are calculated by the field BAB mode selection unit 20, where MCE is an error between the current field BAB and the best candidate field BAB included in the motion compensated field BAB.

In step S6, the MCE is compared with a predetermined reference value TH; If MCE is less than or equal to TH, go to step S7. In step S7, if MVD is zero, the current field BAB is determined to be mode 1, otherwise it is determined to mode 2. In step S8, it is checked whether the mode of the current field BAB is mode 1 or 2. If it is found that mode 1 is assigned to the current field BAB, the mode 1 signal and the FIF generated in step S4 are encoded and terminated in step S9. If it is found that mode 2 is assigned to the current field BAB, the mode 2 signal, FIF and MVD ends in step S10 by transmitting the encoded and encoded data.

Referring to step S6, if MCE is greater than TH, then node B goes to step S11. In step S11, the current field BAB is encoded using the intra-CAE and inter-CAE encoding schemes, the number of bits corresponding to the intra-CAE data and the number of bits corresponding to the inter-CAE data, the intra- shown in FIG. As described for the CAE unit 52-1, the inter-CAE unit 52-2, the intra-bit calculator 54-1, and the inter-bit calculator 54-2, the calculation is performed. If in step S12 the number of intra-CAE bits is less than or equal to the number of inter-CAE bits, go to step S13, otherwise go to step S15.

In step S13, the mode of the current field BAB is set to mode 5, and in step S14, the mode 5 signal is encoded and ends by transmitting the encoded mode signal and intra-CAE data. In step S15, the mode of the current field BAB is determined as mode 6 if MVD is 0, otherwise mode 7 and continues to step S16. In step S16, it is checked whether the mode of the current field BAB is mode 6. In step S16, if it is determined to be mode 6, the mode 6 signal and the FIF are encoded and transmitted together with the inter-CAE data in step S17, and ends. On the other hand, if the mode of the current field BAB is mode 7, the flow goes to step S18. Finally, in step S18, the mode 7 signal, FIF and MVD are encoded and transmitted together with the inter-CAE data to the transmitter, and then ends.

Although the present invention has been described for encoding a binary shape signal on a field-by-field basis, the same scheme can be extended to encode a binary shape signal on a frame-by-frame basis. In this example, the motion estimation and compensation process will be performed for the previous frame instead of the two previous fields.

While specific embodiments of the present invention have been described above, those skilled in the art can make various changes without departing from the scope of the claims described herein.

Claims (7)

In a method of encoding a target block of a binary shape signal, the binary shape signal includes a plurality of pictures, each picture being divided into a plurality of blocks of M × N pixel size having a first and a second binary value. Where the target block is one of the blocks of the current picture to be encoded, and M and N are positive integers: (a) If the error of the target block for the first and second reference blocks is not greater than the predetermined reference value, the mode of the target block is determined as the first and second modes, respectively, where each reference block is M × N pixels. Wherein all pixels of the first and second reference blocks have first and second binary values, respectively; (b) Motion vector information and motion compensation including a motion vector by motion estimation and compensation of the target block for one or more previous pictures for the current picture, if not determined as the first and second modes in step (a). Generating a block, wherein the motion compensation block includes a block most similar to the target block; (c) calculating a motion compensation error (MCE) between the target block and the most similar block and a motion vector difference (MVD) between the motion vector and its estimate; (d) if the MVD is not 0, encoding the MVD to provide encoded MVD data; (e) if the MVD is 0 and MCE is not greater than the reference value, set the mode of the target block to the third mode; if the MVD has a nonzero value and MCE is not greater than the reference value, set the mode of the target block to the fourth mode; ; (f) if not determined as the third or fourth mode in step (e), generate intra-encoded data and inter-encoded data, the intra-encoded data being based on the pixels of the predetermined current picture; Generating pixels by encoding the pixels of the target block, and encoding the pixels of the target block based on predetermined pixels included in the current picture and the motion compensation block; (g) comparing the number of bits of intra-coded data with the number of bits of inter-coded data; (h) if the number of bits of the intra-coded data is not greater than the number of bits of the inter-coded data, setting the mode of the target block to the fifth mode; (i) if it is not determined as the fifth mode in step (h), if the MVD is 0, setting the mode of the target block to the sixth mode; and if the MVD is not 0, setting the mode of the target block to the seventh mode. Binary shape signal encoding method characterized in that. The method of claim 1, wherein the binary shape signal encoding method is followed by step (i): (j) encoding a mode of the target block to generate an encoded mode signal; (k) generating encoded data of the target block according to the mode of the target block. The method of claim 2, wherein the encoded data of the target block includes an encoded mode signal, encoded MVD data in the fourth mode, intra-encoded data in the fifth mode, and in the sixth mode. The inter-encoded data, in the seventh mode, further comprises inter-encoded data and encoded MVD data. In an apparatus for encoding a target block of a binary shape signal, the binary shape signal includes a plurality of pictures, each picture being divided into a plurality of blocks of M × N pixel size having first and second binary values. Where the target block is one of the blocks of the current picture to be encoded, and M and N are positive integers: If the error for the first and second reference blocks of the target block is not greater than the predetermined reference value, first and second type identification signals of type 1 and type 2 are generated for the target block, respectively, where each reference block is M × N pieces. Mode sensing means having pixels and all pixels of the first and second reference blocks each having a first and a second binary value; Reproducing means for reproducing the encoded binary shape signal and for generating and storing the reproduced binary shape signal, wherein the reproduced binary shape signal includes previously coded pictures; The motion vector information and the motion compensation block are generated by motion estimation of the target block on the reproduced binary shape signal corresponding to one or more previous pictures for the current picture, wherein the motion vector information includes the motion vector, and the motion compensation block is Motion estimation and compensation means including a block most similar to the target block; Mode selection means for calculating a motion compensation error (MCE) between the target block and the closest block and generating a second identification signal if the MCE is not greater than a reference value; Motion vector difference (MVD) calculating means for calculating a motion vector difference (MVD) between the motion vector and its estimate and encoding MVD to generate coded MVD data; Intra- and inter-encoding data is performed on the target block based on the target block, the motion compensation block, and the reproduced binary shape signal to generate intra- and inter-coded data, respectively, and the number of bits of the intra- and inter-coded data. Select one of the intra- and inter-encoded data based on the second reference signal; provide a type 3 third identification signal if the intra-encoded data is selected, and provide a type 4 third identification signal if the inter-encoded data is selected Encoding means for performing; And formatting means for determining a mode of the target block and generating coded target block data according to the determined mode according to the MVD, the encoded MVD, the encoded data, and the identification signals. The method of claim 4 wherein the MVD determining means is: Means for storing a motion vector of previously coded blocks in the current picture; Means for finding one of the motion vectors stored in the storage means as an estimate of the motion vector of the target block and calculating the MVD; And means for encoding the MVD to provide coded MVD data. The method of claim 5 wherein the encoding means is: Means for intra-coding the target block based on pixels included in one portion of the reproduced binary shape signal corresponding to the current picture to generate intra-coded data; Means for inter-coding the target block based on the reproduced binary shape signal and the pixels included in one portion of the motion compensation block to provide inter-coded data; Means for finding the number of bits of intra-coded data; Means for finding the sum of the number of bits of the inter coded data and the number of bits of the encoded MVD data; If the sum of the number of bits of the inter coded data and the number of bits of the encoded MVD data is larger than the number of bits of the intra-coded data, a third identification signal of type 3 is generated; otherwise, a third identification signal of type 4 is generated. Means; Means for selecting intra-encoded data as encoded data for a type 3 third identification signal, and means for selecting inter-encoded data as encoded data for a type 3 third identification signal. Encoding device. The method of claim 6 wherein the formatting means: The mode of the target block is determined according to the first to third identification signals and the MVD. If the first identification signal of type 1 is input, the mode of the target block is set to the first mode, and the first identification signal of type 2 is input. In the second mode, when the second identification signal is input and MVD is 0, in the third mode; when the second identification signal is input and the MVD is not 0, in the fourth mode; when the type 3 third identification signal is input, in the fifth mode. Determining means for setting to a sixth mode if a third identification signal of type 4 is input and MVD is 0 and to a seventh mode if a third identification signal of type 4 is input and MVD is not zero; Means for encoding a mode of the target block to provide an encoded mode signal; Generates data of the encoded target block according to the determined mode, the encoded data, and the encoded mode signal, wherein the encoded mode signal is generated in the first, second, or third mode, and the encoded mode signal and the encoded MVD in the fourth mode. Providing means for providing the encoded mode signal and the encoded data in the fifth or sixth mode and the encoded mode signal, the encoded MVD data, and the encoded data in the seventh mode as data of the encoded target block. Binary shape signal encoding apparatus characterized in that.

Images