JP2003102004A - Image coding and decoding method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image coding and decoding method that can efficiently encodes a moving picture signal into a stream including a signal indicative of reference frame selection and a quantization parameter. SOLUTION: A unit converting mode into numerical value MVMap 1 outputs a mode information numeral value ModeVal corresponding to a motion vector MV and a reference frame presence/absence signal HasRef indicative of whether or not a frame referenced by the motion vector MV is a specified value. A multiplication unit Mul1 multiplies a constant K1 with the reference frame presence/absence signal to obtain a correction reference frame presence/absence numeral KMVal. An addition unit Add1 sums the correction reference frame presence/absence numeral and the mode information numeral to output a mode/ reference presence/absence information numeral MVModeVal. When the mode/ reference presence/absence information numeral MVModeVal indicates that the reference frame is a specified value, a selection unit Sel2 outputs no information of a part denoting the reference frame in the motion vector MV as a numeral Va1.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to coding of parameters indicating the presence / absence of inter-picture prediction, a motion compensation unit, and a quantizing step when an image signal is coded using variable length coding in block units. Image encoding method capable of efficiently encoding various information indicating presence / absence, image encoding device, image decoding method for correctly decoding it, image decoding device, and software for implementing the same It is a recording medium in which the program of is recorded.

[0002]

2. Description of the Related Art In recent years, in the era of multimedia in which voices, images, and other pixel values are integrated, information media from the past, that is, newspapers, magazines, televisions, radios, telephones, etc., are transmitted to people. Means have become the subject of multimedia. In general, multimedia means not only characters but also figures, voices, especially images, etc. are associated with each other at the same time. However, in order to target the above-mentioned conventional information media as multimedia, the information is converted into a digital format. It becomes an indispensable condition.

However, when the information amount of 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 that of voice is 64 kbits per second. (Telephone quality), and for moving images, an information amount of 100 Mbits (current television reception quality) or more is required per second, and it is not realistic to handle the enormous amount of information in the digital form as it is in the above information media. For example, videophone is 64kbps ~ 1.5M
Integrated service digital network (IS
Although it has already been put to practical use by DN: Integrated Services Digital Network), it is impossible to send the image of the TV / camera as it is with ISDN.

Therefore, what is needed is an information compression technique. For example, in the case of a videophone, the H.261 and H.263 standards internationally standardized by the ITU (International Telecommunication Union Telecommunication Standardization Sector). The video compression technology of is used.
Further, according to the information compression technology of the MPEG-1 standard, it becomes possible to put image information together with audio information in a normal music CD (compact disc).

Here, MPEG (Moving Picture Experts G
roup) is an international standard for digital compression of moving image signals, and MPEG-1 is a standard for compressing moving image signals up to 1.5 Mbps, that is, information of television signals up to about 1/100. In addition, since the transmission speed for the MPEG-1 standard is mainly limited to about 1.5 Mbps, in MPEG-2 standardized to meet the demand for higher image quality, the moving image signal is 2 to 15 Mbps. Is compressed to.

[0006] Furthermore, at present, a working group that has been promoting standardization with MPEG-1 and MPEG-2 (ISO / IEC JTC1 / SC29 / WG11)
Has standardized MPEG-4, which has a higher compression rate. M
Initially, PEG-4 not only enables highly efficient encoding at a low bit rate, but also introduces powerful error resilience technology that can reduce subjective image quality degradation even if a transmission path error occurs. . In addition, the ITU is promoting the standardization of H.26L as a next-generation image coding method.

FIG. 12 is a block diagram of an image coding apparatus. The motion detection unit ME compares the reference image PrevV read from the memory Mem with the image signal Vin, and detects the motion vector MV that minimizes the inter-screen error. Motion compensation unit
The MC reads the pixel designated by the motion vector MV from the memory Mem and outputs it as the predicted image RefV. Subtraction unit
Sub0 subtracts the predicted image RefV from the image signal Vin and outputs a residual image DifV which is a difference signal. Frequency conversion unit Tr
ans performs frequency conversion such as cosine conversion on the residual image DifV and outputs a pixel coefficient component Coef. The quantization unit Q quantizes the pixel coefficient component Coef with the quantization parameter QP and outputs a quantized value Qcoef. The coefficient information variable length coding unit CoefVLC encodes the pixel coefficient component Coef, and the pixel value coding signal CoefStr and the coding skip information Skip that indicates whether or not a specific part of the pixel coefficient component Coef or all coefficient values are 0. Is output.

The inverse quantization unit IQ is a pixel value encoded signal.
Dequantize CoefStr with a quantization parameter QP and output an inverse quantized value IQCoef. The frequency inverse transform unit ITrans performs frequency inverse transform on the inverse quantized value IQCoef, and the reproduced pixel component ID
Output ifV. Addition unit Add0 is the reproduction pixel component IDifV
And the predicted image RefV are added to obtain a reproduced image RecV, which is stored in the memory Mem for reference in the encoding of the subsequent image.

The encoding mode decision ModeDec is the motion vector M
Refer to V and coding skip information Skip, and set mode information Mode
To decide. Header information variable length coding unit ModeVLC
Encodes the mode information Mode, the motion vector MV and the quantization parameter QP in a variable length,
Output as. The multiplexing unit Mux multiplexes the mode-coded signal ModeStr and the mode-coded signal ModeStr and outputs it as an image-coded signal Str.

In order to limit the number of bits of the image coded signal Str to a certain range, the rate control unit RateCtl monitors the image coded signal Str, and if the number of bits of the image coded signal Str is larger than a predetermined value, the quantum Control by increasing the quantization parameter QP to reduce the number of bits, and if the number of bits of the image coded signal Str is smaller than the specified value, reduce the quantization parameter QP and use more bits to improve the image quality. To improve.

FIG. 13 is a block diagram of a header information variable length coding unit ModeVLC in a conventional image coding apparatus. The motion vector MV, the mode information Mode, and the quantization parameter QP are the mode digitization unit MVMap0, the mode digitization unit ModeMap0, and the mode restoration unit QPIMap0, respectively.
Is the corresponding motion vector value MVVa
Converted to l0, mode information value ModeVal, and quantized value QPVal. The selection unit Sel0 sequentially selects the motion vector numerical value MVVal0, the mode information numerical value ModeVal, and the quantized numerical value QPVal, and outputs it as the numerical value Val. Variable length coding unit VLC
Performs variable length coding of the numerical value Val with a single variable length coding system, and outputs it as a mode coded signal ModeStr.

FIG. 14 is a diagram showing the correspondence between numerical values and codes in variable length coding used in H.26L. When Value; generally represents the numerical value Value + 1 in binary, it is an N-digit binary number b (N
-1), b (N-2), b (N-3), ..., b (2,), b (1), b (0), then 0, b (N-2), 0, b (N-3), 0, b (N-4), 0, ..., b (2), 0, b
(1), 0, b (0), 1 2 × N-1 bits are codewords of Value (Cod
e), and the feature is that the relation between Value and Code can be easily calculated by arithmetic operation without preparing the code table corresponding to all Value.

FIG. 15 is a correspondence diagram of the numerical values and modes of the header information variable length coding unit ModeVLC in the conventional image coding apparatus, and the mode information Mode in the inter-picture coding frame used in Value and H.26L. It shows the relationship of. In H.26L, the difference from the previous image is not coded at all in macro block units (simply abbreviated below as block for the sake of simplicity), other than intra-frame coding (Intra). At 4x4,4x8,8x4,8x8,8x16,16x8,16x16
Prediction error (Inte
r) can be encoded.

FIG. 16 is a diagram showing a stream corresponding to a block of an image coded signal created by a conventional image coding apparatus. Coding mode Coded signal MBTypeStr is shown in Fig. 1.
5 is a coded signal indicating whether the block shown in FIG. 5 is intra-frame coding or inter-frame coding. In H.26L, a reference image can be selected from a plurality of images, and a reference frame number coded signal RefFrameStr is a coded signal that indicates which frame the block refers to. The motion vector coded signal MVStr is a coded signal indicating the motion vector of the block, and the quantization parameter coded signal QPStr is a coded signal representing the quantization parameter (quantization step) of the block. Figure 1
2 and the motion vector MV of FIG. 13 are the reference frame number coded signal RefFrameStr and the motion vector coded signal MVStr.
It is the information that is combined. Subsequent to the mode coded signal ModeStr which is the header of these blocks, the pixel value coded signal CoefStr which is the coded signal of the pixel value of each pixel is arranged.

FIG. 17 is a block diagram of the image decoding apparatus. The separation unit DeMux separates the image coded signal Str into a mode coded signal ModeStr and a pixel value coded signal CoefStr. The header information variable length decoding unit ModeVLD performs variable length decoding of the mode coded signal ModeStr, and the motion vector M
Output V, quantization parameter QP and mode information Mode.
Memory Mem, motion compensation unit MC, inverse quantization unit I
The operation of each unit of Q, the frequency inverse conversion unit ITrans, and the addition unit Add0 is the same as the operation of each unit described in the block diagram of the image encoding device in FIG. In this way, the reproduced image RecV that is the decoded image of the image decoding apparatus is obtained.

FIG. 18 is a block diagram of a header information variable length decoding unit ModeVLD in a conventional image decoding apparatus. Variable length decoding unit VLD is a mode coded signal Mod
The eStr is decoded to convert each codeword forming the mode coded signal ModeStr into a corresponding numerical value Val. The numerical value Val is separated by the selection unit Sel1 into a motion vector numerical value MVVal0, a mode information numerical value ModeVal, and a quantized numerical value QPVal. The mode restoration unit MVIMap0, the mode digitization unit ModeMap0, and the mode restoration unit QPIMap0 are the motion vector numbers M, respectively.
VVal0, mode information value ModeVal, motion vector MV which is a symbol corresponding to quantized value QPVal, mode information Mod
e, Convert to the quantization parameter QP.

[0017]

The conventional image coding apparatus and the conventional image decoding apparatus as described above correspond to the blocks of the image coded signal created by the conventional image coding apparatus of FIG. As shown in the stream, the reference frame number encoded signal RefFrameStr and the quantization parameter encoded signal QPStr must be arranged in the stream.

On the other hand, in MPEG, there is no concept corresponding to the reference frame number coded signal RefFrameStr, and the quantized parameter coded signal QPStr has a defined value. Placement in the stream can be omitted. However, when the symbols of the coding mode coded signal MBTypeStr corresponding to all the mode information Mode including the presence / absence of the quantization parameter coded signal QPStr are prepared, a variable length coding table of enormous size is prepared. Therefore, the arrangement of the quantization parameter coded signal QPStr in the stream can be omitted only in the case of very limited specific mode information Mode.

Therefore, the reference frame number coded signal RefF
The rameStr and the quantized parameter coded signal QPStr often need to be arranged in the stream, which has the drawback of hindering the improvement of the compression rate.

[0020]

In order to solve this problem, a first aspect of the present invention is directed to a first coding step of allocating an integer to a symbol in a block-by-block coding mode and the arithmetic operation of the allocated integer. An image coding method having a variable-length code coded in a second coding step for variable-length coding, comprising α first coding modes and β second coding modes. However, the first coding mode is an integer γ (where γ ≦ α-1) and the second coding mode is an integer δ.
(Where δ ≦ β−1), an image encoding method in which an integer of γ + δ × ε (where ε is a constant ε ≦ α) is an integer to be assigned in the first encoding step.

A second aspect of the present invention comprises a first decoding step of decoding an arithmetically operable variable length code to obtain an integer value and a second decoding step of converting the obtained integer value into a symbol. An image decoding method for encoding, comprising α first encoding modes and β second encoding modes, wherein the second decoding step is γ + δ × ε (where ε is ε Acquiring an integer γ (where γ ≦ α-1) representing the first coding mode and an integer δ (where δ ≦ β-1) representing the second coding mode from an integer value of ≦ α Is the image decoding method.

According to the present invention having the above-mentioned configuration, all the first numbers are represented by performing the variable length coding capable of arithmetically operating the integers represented by the combination of the first coding mode and the second coding mode. Enables coding of a combined representation of the coding mode and the second coding mode. The second coding mode corresponds to the presence / absence of quantization parameter coding and a motion vector coding presence / absence identification signal. The compression rate of arithmetic variable length coding itself is not as high as the method that uses a variable length code table like conventional MPEG, but it can be coded by arithmetic operation without preparing a huge size code table. Therefore, the combined representation of all the first coding modes and the second coding modes can be coded. As a result, the reference frame number coded signal RefFrameStr and the quantization parameter coded signal QPStr often do not need to be arranged in the stream, and the compression rate can be improved as a whole.

Similarly, when an image is decoded, it can be decoded by an arithmetic operation without preparing a code table having an enormous size. Therefore, all the first coding mode and the second coding are possible. The combined representation of modes can be decoded.
As a result, even in the case of all combinations in which the arrangement of the reference frame number encoded signal RefFrameStr and the quantization parameter encoded signal QPStr is omitted in the stream, the image signal can be correctly decoded.

[0024]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to FIGS. 1 to 11.

(Embodiment 1) FIG. 1 shows a header information variable length coding unit ModeVL in the image coding apparatus of the present invention.
FIG. 13 is a block diagram of C and operates as the header information variable length coding unit ModeVLC of FIG. 12. In FIG. 1, units that perform the same operations as those in the block diagram of the image encoding apparatus in FIG.
The description is omitted. That is, the operations of the mode digitizing unit QPMap0 and the variable length coding unit VLC are the same.

The mode digitizing unit MVMap1 outputs a mode information numerical value ModeVal, which is a numerical value corresponding to the motion vector MV, and also a reference frame presence / absence signal HasRef indicating whether or not the frame referred to by the motion vector MV is a prescribed value.
Is output. The constant generator ConstGen1 generates the constant K1,
Constant K1 and reference frame presence / absence signal HasR in multiplication unit Mul1
ef is multiplied to obtain the corrected reference frame presence / absence value KMVal. The mode digitizing unit ModeMap1 outputs a mode information numerical value ModeVal which is a numerical value corresponding to the mode information Mode. Addition unit Add1 has a correction reference frame presence / absence value KMVa
l is added to the mode information numerical value ModeVal, and the mode / reference frame presence / absence information numerical value MVModeVal is output.

The selection unit Sel2 is a motion vector value MVVa.
l1, mode / reference frame presence / absence information numerical value MVModeVal, and quantized numerical value QPVal are sequentially selected and output as numerical values Val respectively. However, mode / reference frame presence / absence information numerical value MVMo
When deVal indicates that the reference frame has a specified value, the information of the portion indicating the reference frame in the motion vector MV is not output as the numerical value Val. Variable length coding unit V
The LC encodes the numerical value Val and outputs it as a mode encoded signal ModeStr.

FIG. 3 shows a stream corresponding to a block of an image coded signal created by the image coding apparatus of the present invention. When the reference frame is not the specified value, as shown in FIG. 3A, the reference frame number encoded signal is the same as in FIG.
RefFrameStr is arranged in the stream, but when the reference frame has a specified value, the reference frame number coded signal RefFrameStr is omitted as shown in FIG. In this way, the coding information of the portion of the motion vector MV indicating the reference frame can be reduced, so that the compression rate can be improved.

FIG. 2 is a diagram showing the correspondence between numerical values and modes of the header information variable length coding unit ModeVLC in the image coding apparatus of the present invention. In FIG. 2, the reference frame number Re
Numerical value of mode information Mode in which fFrame is encoded is
Mode information in which the reference frame number RefFrame is not encoded
The numerical values Value1 to 7 corresponding to the Mode correspond to the numerical values Value9 to 15 obtained by adding 7 respectively. Thus, in the correspondence diagram of FIG. 2, even the parts of the numerical values Value0 to 8 (this part corresponds to the correspondence diagram of the numerical values and modes of the header information variable length coding unit ModeVLC in the conventional image coding device of FIG. 15). If prepared, it is not necessary to store the code table corresponding to the numerical values Value9 to 15, and the memory required for storing the code table can be saved.

In the first embodiment, in the header part, the example in which the motion vector MV, the mode information Mode, and the quantization parameter QP are coded by the header information variable length coding unit ModeVLC has been described. Header information variable-length coding unit Mode because it is encoded by other means.
When it is not necessary to encode with VLC, only the motion vector MV and mode information Mode are header information variable length encoding unit.
It may be encoded by ModeVLC.

(Second Embodiment) FIG. 4 shows a header information variable length decoding unit ModeVL in the image decoding apparatus of the present invention.
It is a block diagram of D. FIG. 4 operates as the header information variable length decoding unit ModeVLD of FIG. 17, and performs the reverse operation of the header information variable length coding unit ModeVLC of FIG.
In FIG. 4, units that perform the same operations and signals that perform the same operations as those in the block diagram of the header information variable length decoding unit ModeVLD in the conventional image decoding apparatus in FIG. 18 are given the same symbols, and description thereof will be omitted. That is, the operations of the variable length decoding unit VLD and the mode restoration unit QPIMap0 are the same.

The numerical value Val is classified by the selection unit Sel3 into one of a motion vector numerical value MVVal1, a mode / reference frame presence / absence information numerical value MVModeVal, and a quantized numerical value QPVal. The constant generator ConstGen1 generates a constant K1, and the division unit Div1 calculates the mode / reference frame presence / absence information numerical value MVModeVa
Divide l by the constant K1. The quotient divided by the constant K1 becomes the reference frame presence / absence signal HasRef, and the remainder divided by the constant K1 becomes the mode information numerical value ModeVal. However, if only some modes of the code table can be arithmetically configured, this division procedure is applied only to the part that can be arithmetically configured.

For example, in the diagram of the correspondence between the numerical values and modes of the header information variable length coding unit ModeVLC in the image coding apparatus of the present invention shown in FIG. 2, the mode information Mode of numerical values Value1 to 7 has the configuration of numerical values Value9 to 15. Since it is used,
If the constant K1 is 8, the numeric value is not regular except 0 and 8, the quotient obtained by dividing the numeric value by the constant K1 is the reference frame number RefFrame, and the remainder when the numeric value is divided by the constant K1 is the mode information Mode. .

The mode restoration unit MVIMap1 refers to the reference frame presence / absence signal HasRef, and if the reference frame is not coded, restores the motion vector value MVVal1 by referring to the default reference frame to obtain the motion vector MV, When the reference frame is encoded, the motion vector MV including information indicating the reference frame is restored from the motion vector value MVVal1.

As described above, according to the present embodiment, the image coded signal Str generated by the image coding apparatus using the header information variable length coding unit ModeVLC in FIG. 1 can be correctly decoded. it can.

In the first embodiment, in the header section, an example in which the motion vector MV, the mode information Mode, and the quantization parameter QP are decoded by the header information variable length coding unit ModeVLC has been described. Header information variable length coding unit Mode because it is decoded by other means.
When there is no need to decode with VLC, only the motion vector MV and mode information Mode are header information variable length coding unit.
You may decrypt with ModeVLC.

(Third Embodiment) FIG. 5 shows a header information variable length coding unit ModeVL in the image coding apparatus of the present invention.
FIG. 13 is a block diagram of C and operates as the header information variable length coding unit ModeVLC of FIG. 12. 5, units having the same operation as those in the block diagram of the image coding apparatus in FIG. 12 and signals having the same operation are denoted by the same symbols,
The description is omitted. That is, the operations of the mode digitization unit MVMap0 and the variable length coding unit VLC are the same.

The mode digitization unit QPMap1 outputs a quantized value QPVal, which is a number corresponding to the quantized parameter QP, and outputs a quantized parameter coding presence / absence signal HasQP indicating whether or not the quantized parameter QP is a prescribed value. Output. The constant generator ConstGen2 generates the constant K2, and the multiplication unit Mul2 outputs the constant K2 and the quantization parameter coding presence / absence signal.
HasQP multiplication is performed, and the numerical value KQP with or without correction quantization parameter
Get Val. The mode digitizing unit ModeMap1 outputs a mode information numerical value ModeVal which is a numerical value corresponding to the mode information Mode. The addition unit Add2 adds the correction quantization parameter presence / absence value KQPVal and the mode information value ModeVal, and outputs a mode / quantization parameter presence / absence information value QPModeVal.

The selection unit Sel4 is a motion vector value MVVa.
l0, mode / quantization parameter presence / absence information Numerical value QPModeVa
l and the quantized numerical value QPVal are sequentially selected and output as numerical values Val respectively. However, when the mode / quantization parameter presence / absence information numerical value QPModeVal indicates that the quantization parameter is a specified value, the quantized numerical value QPVal of the block is not output as the numerical value Val. The variable length coding unit VLC codes the numerical value Val and outputs it as a mode coded signal ModeStr.

FIG. 7 is a diagram showing a stream corresponding to a block of an image coded signal created by the image coding apparatus of the present invention. If the quantization parameter is not a specified value,
As shown in FIG. 6 (a), the quantization parameter coded signal QPStr is arranged in the stream as in FIG. 16, but if the reference frame has a specified value, quantization is performed as shown in FIG. 7 (b). The parameter coded signal QPStr is omitted. In this way, when the quantization parameter is the specified value, the coding information of the quantization parameter coded signal QPStr can be reduced, so that the compression rate can be improved.

FIG. 6 is a diagram showing the correspondence between numerical values and modes of the header information variable length coding unit ModeVLC in the image coding apparatus of the present invention. In FIG. 6, the quantization parameter QP
Numerical value of the mode information Mode in which is encoded is the numerical value Valu obtained by adding 8 to each of the numerical values Value1 to 8 corresponding to the mode information Mode in which the quantization parameter QP is not encoded.
e9-16 will be supported. Thus, in the correspondence diagram of FIG. 6, even the parts of the numerical values Value0 to 8 (this part corresponds to the correspondence diagram of the numerical values and modes of the header information variable length coding unit ModeVLC in the conventional image coding device of FIG. 15) If prepared, it is not necessary to store the code table corresponding to the numeric values Value 9 to 16, and the memory required for storing the code table can be saved.

In the third embodiment, in the header section, the example in which the motion vector MV, the mode information Mode, and the quantization parameter QP are coded by the header information variable length coding unit ModeVLC has been shown. Header information variable length coding unit ModeVLC because it is coded by other means
If it is not necessary to encode with, only the quantization parameter QP and mode information Mode are header information variable length coding unit.
It may be encoded by ModeVLC.

(Embodiment 4) FIG. 8 shows a header information variable length decoding unit ModeVL in the image decoding apparatus of the present invention.
It is a block diagram of D. FIG. 8 operates as the header information variable length decoding unit ModeVLD of FIG. 17, and performs the reverse operation of the block diagram of the header information variable length coding unit ModeVLC of the image coding apparatus of the present invention of FIG. In FIG. 8, units that perform the same operations and signals that perform the same operations as those in the block diagram of the header information variable length decoding unit ModeVLD in the conventional image decoding apparatus of FIG. 18 are given the same symbols, and description thereof is omitted. That is, the variable length decoding unit VLD
And the operation of the mode restoration unit MVIMap0 is the same.

The numerical value Val is classified by the selection unit Sel5 into one of a motion vector numerical value MVVal0, a mode / quantization parameter presence / absence information numerical value QPModeVal, and a quantized numerical value QPVal. The constant generator ConstGen2 generates a constant K2, and the division unit Div2 outputs the mode / quantization parameter presence / absence information numerical value QP.
Divide ModeVal by the constant K2. The quotient divided by the constant K2 becomes the quantization parameter coding presence / absence signal HasQP, and the remainder divided by the constant K2 becomes the mode information numerical value ModeVal. However,
If only some modes of the codebook can be configured arithmetically, then this division procedure is applied only to the part that can be configured arithmetically.

For example, in the correspondence diagram of the numerical values and modes of the header information variable length coding unit ModeVLC in the image coding apparatus of the present invention shown in FIG. 6, the mode information Mode of numerical values Value1 to 8 has the configuration of numerical values Value9 to 16. Since it is used,
If the constant K2 is set to 8 and the numerical Value is not regular and 0 is excluded, if 1 is subtracted from the numerical value beforehand, the quotient obtained by dividing the numerical value Value-1 by the constant K2 indicates whether or not the quantization parameter QP is encoded.
Mode information M with the same remainder when the numerical value Value-1 is divided by the constant K2
It becomes a numerical value indicating ode.

The mode restoration unit QPIMap1 refers to the quantization parameter coding presence / absence signal HasQP, and when the quantization parameter is not coded, obtains the quantization parameter QP as a default quantization parameter, and quantizes it. If the parameter is encoded, the quantization parameter QP is restored from the quantized value QPVal.

As described above, according to the present embodiment, the image coded signal Str generated by the image coding apparatus using the header information variable length coding unit ModeVLC of FIG. 1 can be correctly decoded. it can.

In the fourth embodiment, the example in which the motion vector MV, the mode information Mode and the quantization parameter QP are coded by the header information variable length coding unit ModeVLC in the header part has been described. Header information variable length coding unit ModeVLC because it is decoded by other means
If there is no need to decode in, only the quantization parameter QP and the mode information Mode are included in the header information variable length coding unit.
You may decrypt with ModeVLC.

(Embodiment 5) A combination of Embodiments 1 and 3, or Embodiments 2 and 4, combining with or without quantization parameter QP coding, and reference frame number RefF
The presence / absence of rame encoding may be collectively encoded as mode information.

FIG. 9 is a correspondence diagram of numerical values and modes of the header information variable length coding unit ModeVLC in the image coding apparatus of the present invention. In the example of FIG. 9, quantization parameter QP coding is performed, reference frame number RefFrame coding is performed,
The corresponding mode can be calculated arithmetically according to the following rules.

(When Numerical Value is Generated) 1) When the quantization parameter QP and the reference frame number RefFrame are not coded, the value shown in FIG. 15 is used. 2) When only the quantization parameter QP is coded Value is shown in FIG. 15: Add 8 to the same Mode 3) When only the reference frame number RefFrame is encoded Value: Add 16 to the same Mode 4 in Fig. 15 4) Quantization parameter QP and reference frame number RefFrame are encoded If the quotient obtained by dividing Value by 16 is 1, there is a reference frame number encoded signal RefFrameStr, and if 0, the reference frame number encoded signal is present. RefFrameStr None 1) When Value is less than 17 If the quotient obtained by dividing Value-1 by 8 is 1, there is a quantization parameter coded signal QPStr, and if it is 0, there is a quantization parameter coded signal QP.
Str None Value-1 is divided by 8 and 1 is added to the remainder, which corresponds to the mode in Fig. 15 2) When Value is less than 17, if Value-3 is divided by 7 and the quotient is 3, quantization parameter encoding Signal QPStr present, if 2 then quantization parameter coded signal QP
Str None Value-3 corresponding to the value obtained by adding 1 to the remainder obtained by dividing by 3 and the mode coded signal corresponding to 1), 2), 3), 4) when the mode of FIG. Examples of ModeStr are shown in FIGS.

(Sixth Embodiment) Furthermore, a program for realizing the configuration of the image coding method or the image decoding method shown in each of the above embodiments is recorded in a storage medium such as a floppy (registered trademark) disk. By doing so, it becomes possible to easily carry out the processing shown in each of the above embodiments in an independent computer system.

FIG. 11 is an explanatory diagram of a case where the present invention is carried out by a computer system using a floppy disk storing the image coding method or the image decoding method of the first to second embodiments.

FIG. 11B shows the external appearance, sectional structure, and floppy disk of the floppy disk as viewed from the front, and FIG. 11A shows an example of the physical format of the floppy disk, which is the recording medium body. . The floppy disk FD is built in a case F, and a plurality of tracks Tr are formed concentrically from the outer circumference toward the inner circumference on the surface of the disk F, and each track is formed in an angular direction by one.
It is divided into 6 sectors Se. Therefore, in the floppy disk storing the program, the image encoding method and the image decoding method as the program are recorded in the area allocated on the floppy disk FD.

FIG. 11C shows a structure for recording / reproducing the above-mentioned program on the floppy disk FD. When the program is recorded on the floppy disk FD, the image encoding method or the image decoding method as the program is written from the computer system Cs via the floppy disk drive. When the image coding method and the image decoding method are constructed in a computer system by a program in a floppy disk, the program is read from the floppy disk by a floppy disk drive and transferred to the computer system.

In the above description, a floppy disk is used as the recording medium, but an optical disk may be used. Further, the recording medium is not limited to this, and any recording medium such as an IC card and a ROM cassette that can record the program can be similarly used.

[0057]

As described above, according to the image coding method and the image decoding method according to the present invention, the combined representation of all the coding modes is coded without preparing a code table of enormous size. Can be converted. As a result, the reference frame number coded signal RefFrameStr and the quantization parameter coded signal QPStr often do not need to be arranged in the stream, and as a whole, the compression rate can be improved and the practical value is high.

[Brief description of drawings]

FIG. 1 is a block diagram of a header information variable length coding unit ModeVLC in an image coding apparatus according to a first embodiment of the present invention.

FIG. 2 is a diagram showing correspondence between numerical values and modes of a header information variable length coding unit ModeVLC in the image coding device according to the first embodiment of the present invention.

FIG. 3 is a diagram showing streams corresponding to blocks of an image coded signal created by the image coding apparatus according to the first embodiment of the present invention.

FIG. 4 is a block diagram of a header information variable length decoding unit ModeVLD in an image decoding device according to a second embodiment of the present invention.

FIG. 5 is a block diagram of a header information variable length coding unit ModeVLC in the image coding apparatus according to the third embodiment of the present invention.

FIG. 6 is a diagram showing correspondence between numerical values and modes of a header information variable length coding unit ModeVLC in the image coding apparatus according to the third embodiment of the present invention.

FIG. 7 is a diagram showing streams corresponding to blocks of an image coded signal created by the image coding apparatus according to the third embodiment of the present invention.

FIG. 8 is a block diagram of a header information variable length decoding unit ModeVLD in an image decoding apparatus according to a fourth embodiment of the present invention.

FIG. 9 is a diagram showing correspondence between numerical values and modes of the header information variable length coding unit ModeVLC in the image coding apparatus according to the fifth embodiment of the present invention.

FIG. 10 is a diagram showing a stream of an image coded signal created by the image coding apparatus according to the fifth embodiment of the present invention.

FIG. 11 is an explanatory diagram of a storage medium for storing a program for implementing the image coding method and the image decoding method according to the sixth embodiment of the present invention by a computer system.

FIG. 12 is a block diagram of an image encoding device.

FIG. 13 is a block diagram of a header information variable length coding unit ModeVLC in a conventional image coding apparatus.

FIG. 14 is a diagram showing correspondence between numerical values and codes in variable-length coding.

FIG. 15 is a diagram showing correspondence between numerical values and modes of a header information variable length coding unit ModeVLC in a conventional image coding apparatus.

FIG. 16 is a diagram showing a stream corresponding to a block of an image coded signal created by a conventional image coding apparatus.

FIG. 17 is a block diagram of an image decoding device.

FIG. 18 is a block diagram of a header information variable length decoding unit ModeVLD in a conventional image decoding device.

[Explanation of symbols]

Vin Video signal Vout Decoded video signal Str Image coded signal ME Motion detection unit MC Motion compensation unit Mem Memory Trans Frequency conversion unit Itrans Frequency inverse transform unit Q Quantization unit IQ inverse quantization unit Mux Multiplexing unit DeMux Separation unit CoefVLC Coefficient information variable length coding unit CoefVLD Coefficient information variable length decoding unit RateCtl Rate control unit ModeDec Coding mode decision ConstGen1, ConstGen2 Constant generator MVMap0, MVMap1, ModeMap0, ModeMap1, QPMap0, QPMap1
Mode digitization unit MVIMap0, MVIMap1, ModeIMap0, ModeIMap1, QPIMap0, Q
PIMap1 Mode restoration unit ModeVLC Header information variable length coding unit ModeVLD Header information variable length decoding unit Sel0, Sel1, Sel2, Sel3, Sel4, Sel5 Selection unit Mul1, Mul2 Multiplication unit Add0, Add1, Add2 Addition unit Sub0 Subtraction unit Div1, Div2 Division unit VLC Variable length coding unit VLD Variable length decoding unit Cs Computer system FD floppy disk FDD floppy disk drive

   ─────────────────────────────────────────────────── ─── Continued front page    F-term (reference) 5C059 MA27 MD02 ME01 ME11 ME17                       NN01 PP01 PP04 RB02 RB09                       RB12 RC32 RC37 SS20 UA02                       UA38 UA39                 5J064 AA02 BA09 BB03 BB06 BC01                       BC08 BC09 BC16

Claims (10)

[Claims] 1. A variable length coded block-by-block coding mode in a first coding step of allocating integers to symbols and a second coding step of variable-length coding of the allocated integers capable of arithmetic operation. An image coding method having a code, which has α first coding modes and β second coding modes, where the first coding mode is an integer γ (where γ ≦ α-1 ), If the second coding mode is represented by an integer δ (where δ ≦ β-1), then γ + δ × ε (where ε
Is a constant that satisfies ε ≦ α), which is an integer assigned in the first encoding step. 2. A block-based image coding method, wherein an identification signal indicating whether or not a signal specifying an image to be referred to as a reference image is coded is the second coding mode. The described image coding method. 3. A block-based image coding method, wherein the identification signal indicating whether or not the parameter representing the quantization step of the block is coded is the second coding mode. Image coding method. 4. An image decoding for decoding a variable length code capable of arithmetic operation to obtain an integer value by a first decoding step and a second decoding step for converting the obtained integer value into a symbol. A first encoding mode in α ways and a second encoding mode in β ways, wherein the second decoding step is γ + δ × ε (where ε is a constant such that ε ≦ α). ), An integer γ representing the first encoding mode (where γ ≦ α-1) and an integer δ representing the second encoding mode.
An image decoding method for acquiring (where δ ≦ β-1). 5. A block-by-block image decoding method, wherein an identification signal indicating whether or not a signal specifying an image to be referred to as a reference image is encoded is the second encoding mode. The described image decoding method. 6. A block-by-block image decoding method, wherein the identification signal indicating whether or not the parameter representing the quantization step of the block is coded is the second coding mode. Image coding method. 7. A block-unit coding mode is provided with a first coding means for allocating an integer to a symbol and a second coding means for coding the allocated integer, and the first coding means has α patterns. Of the first coding mode and β second coding modes, where the first coding mode is an integer γ (where γ ≦ α-1) and the second coding mode is an integer δ (where δ ≦
γ + δ × ε (where ε is ε ≤ α when expressed as β-1)
Variable constant encoding device for image encoding, wherein the integer is a constant that is assigned by the first encoding means. 8. A first decoding means for decoding an integer value by decoding with a variable length code, and a second decoding means for converting the decoded integer value into a symbol, the second decoding means. The decoding means uses an integer value of γ + δ × ε (where ε is a constant such that ε ≦ α) and an integer γ representing the first encoding mode (where γ
≤α-1) and an integer δ representing the second coding mode (where δ≤
β-1) variable length decoding device for image decoding. 9. A storage medium for storing a program for performing the image coding method according to claim 1 by a computer, wherein the program assigns an integer to a symbol as a block-based coding mode for the computer. An image coding method having a variable length code which is coded in a coding step of 1 and a second coding step of performing a variable length coding capable of arithmetically operating the allocated integers,
It has α first coding modes and β second coding modes, and the first coding mode is an integer γ (where γ ≦ α
-1), and when the second coding mode is expressed by an integer δ (where δ ≦ β-1), an integer of γ + δ × ε (where ε is a constant such that ε ≦ α) is used for the first coding. A storage medium for performing an image coding method, which is an integer assigned in steps. 10. A storage medium for storing a program for performing the image decoding method according to claim 5 by a computer, wherein the program decodes an arithmetically operable variable length code into an integer value. An image decoding method for performing decoding in a first decoding step of obtaining a first encoding mode and a second decoding step of converting the obtained integer value into a symbol, the first encoding mode having α ways and the first encoding mode having β ways In the second decoding step, and in the second decoding step, an integer γ representing the first coding mode (where γ is γ + δ × ε (where ε is a constant ε ≦ α)) ≤ α
-1) and an integer δ representing the second coding mode (where δ ≦ β-
A storage medium characterized in that the image decoding method for obtaining 1) is performed.