JPH06253282A - Image signal encoding / decoding device - Google Patents

Abstract

PURPOSE:To enhance the compression rate by generating user data to add a coded bit stream, setting a decoding parameter based on a coded parameter and user data so as to allow the device to hardly receive the effect of a code error caused in a storage medium. CONSTITUTION:A user data generating section 3 adds error check code and an error correction code to a coded parameter CP or applies shuffling processing to the parameter to generate user data UD. A video multiplexer section 4 applies prescribed time series division multiplexing to data VE, CP, UD to generate coded bit stream data CB. The data are inputted to a transmission buffer section 5. A transmission line coding section 6 generates data VS adding an error correction code to output data from a transmission buffer section 5 to correct a code error caused in a transmission storage medium 7. Then the data VS are communicated and recorded in the transmission storage medium 7. A decoding section uses a transmission line decoding section 8 to apply correction processing of code error caused in the transmission storage medium 7. Then the data subjected to correction processing are inputted to a reception buffer section 9.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital image signal coding / decoding device, and more particularly, it is suitable for efficiently compressing and transmitting / accumulating an information amount by signal processing of high efficiency coding. The present invention relates to an image signal encoding / decoding device.

[0002]

2. Description of the Related Art A digitized image signal has an enormous amount of information, and if it is directly communicated via a transmission line or stored in a recording medium, there are many problems in terms of cost and speed. For this reason, high-efficiency encoding signal processing that efficiently compresses the amount of information is performed by utilizing image redundancy, and images are transmitted / stored at an extremely low bit rate of several tens of kilobits / second to several megabits / second. Research and development of technology is in progress.

In transmission and storage of images at these low bit rates, in many cases, motion compensation predictive coding, DCT
(Discrete Cosine Transform) The signal amount of high efficiency coding that combines orthogonal transform coding, entropy coding, etc., compresses the amount of information to several tenths. For example, CCITT Recommendation H.264. H.261 video encoding or ISO / IECJ
In the moving picture coding MPEG for storage media of TC1 / SC29 / WG11, etc., high-efficiency coding is performed on the image of each frame hierarchically configured, and the obtained coded data and coding parameters (coding Mode, quantization scale, encoded data of DCT coefficient, etc.) are transmitted and stored. Then, the decoding unit decodes the coded data by a predetermined decoding process according to the coding parameter, and reproduces the original image signal. Then, image transmission / storage at a low bit rate is realized.

[0004]

In the transmission / accumulation of the low bit rate image described above, the reproduction image quality is significantly deteriorated due to a code error generated in the transmission / accumulation medium. In order to avoid this, a code error countermeasure such as adding an error correction code is adopted. However, the degree of image quality deterioration due to a code error is greatly different between the coding parameter section and the coded data section. When a code error occurs in the coding parameter part, the block corresponding to this cannot be decoded. For example, even when there is no code error in the coded data part, the entire block is converted into a correlated image signal. It is necessary to perform error correction for replacement. Therefore, the image quality is greatly deteriorated. On the other hand, in the case of a code error occurring in the encoded data portion, the signal up to immediately before this can be decoded, so that the error correction area is narrow and the degree of image quality deterioration is small. However, in the conventional code error countermeasures, no consideration is given to the difference in the degree of image quality deterioration due to the above-mentioned code error, and there is a problem in that it is easily affected by the code error generated in the transmission / storage medium.

It is an object of the present invention to provide an image signal encoding / decoding device which is not easily affected by a code error occurring in a transmission / storage medium and which is excellent in compression efficiency of an information amount.

[0006]

In order to achieve the above object, according to the present invention, a coding unit separately generates user data from coding parameter data for high efficiency coding and adds the user data to a coded bit stream, The decoding unit sets the coding parameter of the signal processing for decoding based on the data of the coding parameter and the user data, and decodes the coded data.

Further, the user data is constructed in a form capable of error detection or error correction with respect to a code error occurring in the process of transmission / accumulation. Further, the user data is shuffled within a specific period of the encoded bit stream to be dispersed.

[0008]

In the present invention, the information on the coding parameters for high efficiency coding is transmitted / stored as user data in addition to the same coding parameter data as in the prior art. Therefore, even if a coding error occurs in the coding parameter data in the transmission / accumulation medium, the information of the regular coding parameter can be decoded by the information of the coding parameter of the user data. Then, the encoded data can be decoded into the original image signal by the decoded encoding parameter. That is, in the conventional technique, even for a coding error of coding parameter data that cannot be decoded into an image signal, in the present invention, a normal coding parameter without a coding error can be decoded by the information of the user data, and the original image It becomes possible to decode into a signal. Further, the information newly added as user data can be realized with almost negligible amount in comparison with the total number of encoded bits.

The user data shuffling process can be effectively operated against a burst error caused by, for example, cell loss in the B-ISDN network or dropout in the storage medium.

As described above, according to the present invention, it is possible to realize an image signal encoding / decoding device which is not easily affected by a code error occurring in a transmission / storage medium and which is excellent in compression efficiency of an information amount. Will be possible.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described with reference to the block diagram shown in FIG.

The digitized image signal VI is input to the information source coding unit 1 and is subjected to predetermined signal processing such as high efficiency coding, such as motion compensation predictive coding, DCT orthogonal transform coding and entropy coding. Video coding is performed to generate coded data VE. On the other hand, the encoding control unit 2 controls, according to the buffer capacity information BF of the transmission buffer unit 5, encoding mode, quantization scale, encoded data of DCT coefficients, and the like necessary for video encoding. The signal CT and the predetermined encoded parameter data CP corresponding thereto are generated. Further, the user data generation unit 3 adds the error detection code and the error correction code to the coded parameter data CP, or performs shuffling processing to generate the user data UD. Then, in the video multiplex unit 4, the data VE, CP, U
D is subjected to predetermined time series division multiplexing to generate coded bit stream data CB. Then, this data is input to the transmission buffer unit 5. In the transmission path coding unit 6, data obtained by adding an error correction code (for example, BCH code, Reed-Solomon code, etc.) to the output data of the transmission buffer unit 5 in order to correct a code error generated in the transmission / storage medium 7. Generate VS. Then, this data VS is communicated and recorded on the transmission / storage medium 7.

On the other hand, in the decoding unit, first, the transmission line decoding unit 8 corrects the code error generated in the transmission / storage medium. Then, the corrected data is input to the reception buffer unit 9. In the video multiplex section 10,
The output data of the reception buffer unit 9 is separated into encoded data VE, encoded parameter data CP ', and user data UD, respectively. The user data decoding unit 11 decodes the coded parameter data CP required for video decoding based on the data CP 'and UD. And
The decoding control unit 12 generates a decoding control signal CTD required for video decoding according to the coding parameter data CP. In the information source decoding unit 13, the decoding control signal C
Predetermined decoding signal processing is performed based on TD to decode the original digital image signal VO.

In the following, ISO / IEC JTC1 / S
C29 / WG11 moving image standard MPEG for storage media
Each block part will be described by taking a video coding conforming to the above as an example.

First, an outline of MPEG video coding will be described with reference to FIG. The figure (a) shows an encoding structure. In order to realize the random access function, GO is used in MPEG.
A layer composed of a plurality of image frames called P (Group of Pictures) is provided. The GOP is composed of three types of coded frames, that is, an intra-frame coded frame (abbreviated as I picture), an inter-frame coded frame (abbreviated as P picture), and a frame interpolated coded frame (abbreviated as B picture). It consists of. Then, in P and B pictures, interframe prediction is performed using reference frames as shown in (b) and (c) of the same figure, and the information amount is efficiently compressed.

On the other hand, each frame (picture) is hierarchically composed of slice, macroblock, and block layers. A picture is composed of one or more slices, a slice is one or more macroblocks, and a macroblock is composed of six blocks (four luminance signal Y blocks and two color difference signals C _B and C _R blocks). .
The block size is 8 pixels × 8 lines.

As a coding algorithm, a hybrid coding method using motion-compensated interframe prediction and DCT orthogonal transform is adopted.

FIG. 3 shows an outline of the structure of a coded bit stream in MPEG. The coded bitstream is hierarchically configured from the video sequence layer to the macroblock layer. Each layer has a header part (S in the figure)
EQ HD, GOP HD, ..., MB HD area)
And encoded data. The header part is as shown in FIG.
There is a coding parameter area, and parameter data CP for video coding is assigned. For example, the frame size, the pixel aspect ratio, the frame frequency, the quantization matrix in the video sequence layer, the time code in the GOP layer, and the coding mode (I, P, B) in the picture layer.
The type of picture), the motion vector size, the quantization scale in the slice layer, the prediction type, the motion vector in the macroblock layer, the encoded data of the DCT coefficient, and other information data are assigned. In addition, the header portion has a user data area in which the user can enter extended data and user data UD. In the present invention, data generated based on the encoded parameter data is inserted as user data in this user data area.

FIG. 5 shows a configuration example of user data in the present invention. FIG. 9A shows that user data is formed by adding an error correction code. That is, the coding parameter data CP is used as the data of the information point of the error correction code, and the data of the check point necessary for error correction (ECC parity in the figure) is added to form the user data. The decoding unit corrects the code error by calculating the syndrome.
On the other hand, FIG. 2B shows a plurality of identical encoded parameter data arranged side by side to form user data. Then, the decoding unit decodes the coding parameter having no code error by a majority decision or the like.

FIG. 6 shows an example of the data structure of the user data shuffling process according to the present invention. In this example, the user data UD (i-1) generated based on the coding parameter data CP (i) (i = 1, 2, 3, 4) of the macroblock MB (i) is used as the header portion MB (i -1)
To the user data area to realize the shuffling process. That is, in the header portion MB (1), MB (1)
Data UD generated with the coding parameter CP (1) of MB and the coding parameter CP (2) of MB (2)
Allocate the data of (1). Therefore, even if a burst error occurs in the header part due to the shuffling process that allocates the encoded parameter data and the user data generated based on this data to different header parts, it is possible to It becomes possible to correctly decode the coding parameters required for video decoding.

Next, the information source coding unit 1 will be described with reference to an embodiment shown in FIG. When the coding mode is the I picture according to the coding control signal CT, the switch unit 26 is connected to the terminal a and performs intraframe coding. When the coding mode is P picture, the switch unit 26 is connected to the terminal b and the switch unit 23 is connected to the terminal a to perform interframe coding. On the other hand, in the B picture, the switch unit 23 is connected to the terminal a and the switch unit 26 is connected to any of the terminals a, b, and c to perform frame interpolation coding.

The digitalized image signal VI is input to the frame memory 14. In the motion vector detection unit 16,
The motion vector information V required for motion compensation predictive coding is extracted. The subtraction circuit 16 subtracts the output signal of the switch unit 26 from the output of the frame memory 14 to calculate the prediction error signal P.
Extract E. The DCT calculator 17 performs a calculation with a predetermined DCT transformation matrix of 8 pixels × 8 lines to calculate a DCT coefficient. Then, the quantizer 18 quantizes the DCT coefficient on the quantization scale determined by the coding control signal CT. The Huffman coding unit 19 performs variable coding in which a code having a short code length is assigned to a code having a high occurrence probability,
Encoded data VE is generated. Also, the inverse quantizer 20,
The inverse DCT calculator 21 performs inverse quantization and DCT transform inverse matrix calculation, respectively, and reproduces the prediction error signal PE ′. Then, the adder circuit 22 adds it to the output signal of the frame memory 14 and decodes it into the image signal of the original frame. The forward prediction value generation unit 24 generates a prediction signal that is motion-compensated for encoding P and B pictures based on the motion vector information V. On the other hand, the backward prediction value generation unit 25
Then, a prediction signal that has been motion-compensated for encoding B pictures is generated.

Next, a first embodiment of the user data generator 3 will be described with reference to FIG. This is suitable for generating the user data shown in FIG. The encoded parameter data CP is input to the memory unit 27. EC
In the C coding unit 28, the generator polynomial G of the error correction code G
Data of the inspection point is generated by (X), and this is added to the encoded parameter data to generate the user data UD. Various codes such as Hamming code and BCH code can be applied to the error correction code. In addition, the shuffling process of user data as shown in FIG.
This can be easily realized by controlling the order of reading the encoded parameter data from 7.

FIG. 9 shows a configuration example of the user data decoding unit 11 of the decoding unit corresponding to this embodiment. User data U
D is input to the memory circuit 27 and is subjected to deshuffling processing or the like which is the reverse of that of the encoding unit. The ECC decoding unit 27 generates coded parameter data CP ″ by calculating the syndrome by a predetermined calculation and correcting the code error. Further, when a code error that cannot be error-corrected is included, The error flag EF is output, and the selection unit 30 connects to the terminal b if there is no error flag EF and connects to the data CP ″, and if there is the error flag EF, connects to the terminal a to connect to the data C.
P '(data in the coding parameter area) is selected and the coding parameter data CP necessary for video decoding is reproduced.

Next, an embodiment of the user data generator 3 suitable for generating the user data shown in FIG. 5B will be described with reference to FIG.
Shown in. The encoded parameter data CP is input to the memory unit 31. Then, the signal reading operation from the memory unit 31 is continuously performed a plurality of times to generate the user data UD. Note that the shuffling process of the user data can be realized at the same time by controlling the order of reading the encoded parameter data.

FIG. 11 is a configuration example of the user data decoding unit 11 of the decoding unit corresponding to this embodiment. This user data UD is input to the memory circuit 31 and is subjected to deshuffling processing, which is the reverse of the encoding process. In the coincidence determination section 32, the encoded parameter data CP ″ of the user data section,
And the majority decision processing is performed on the data CP ′ in the coding parameter area, and a code having many identical patterns is output as the coding parameter data CP. The majority decision processing can be performed in various forms from a bit unit to an encoding parameter data unit.

Next, an embodiment of the information source decoding unit 13 will be described with reference to FIG. By the decoding control signal CTD,
When the coding mode is I picture, the switch unit 36 is the terminal a, when it is the P picture, the switch unit 36 is the terminal b, when switch 41 is the terminal a, when it is B picture, the switch unit 36 is the terminal b, The switch unit 41 has terminals a, b,
The video decoding processing is performed by connecting to any one of c and the digitized image signal VO is decoded.

The encoded data VE is sent to the Huffman decoding unit 3
3 is converted to the original fixed-length code. The inverse quantum unit 34 performs the inverse quantization processing with the characteristics of the quantization scale determined by the control signal CTD to decode the DCT coefficient. The inverse DCT calculator 35 calculates the DCT coefficient and the inverse DCT matrix and decodes the prediction error signal PE ′. The adding circuit 37 adds the prediction error signal PE ′ to the motion-compensated prediction signal obtained from the output of the switch unit 41 to decode the original image signal. In the forward prediction value generation unit 39,
The motion vector information V obtained from the control signal CTD is used to perform motion compensation signal processing on the output signal of the frame memory 38 to generate a prediction signal used for decoding P and B pictures. Further, the backward prediction value generation unit 40 similarly uses the motion vector information V to generate a prediction signal used for decoding a B picture.

As described above, according to the present embodiment, the moving image standard MP for storage media, which is not easily affected by the code error generated in the transmission / storage medium, and which is excellent in the compression efficiency of the information amount.
Image signal coding for video coding based on EG
A decoding device can be realized.

Next, the present invention will be described in accordance with CCITT Recommendation H.264. 26
An embodiment applied to video coding conforming to 1 will be described. It should be noted that this overall block configuration can be configured similarly to that shown in FIG. Also, except that there is no coding type corresponding to the B picture of MPEG, the coding configuration and the coding algorithm are very similar. Therefore, the description regarding these is omitted.

FIG. 13 shows an embodiment of the information source coding unit 1. When the coding type is intra-frame coding, the switch unit 42 is connected to the terminal a, and when it is inter-frame coding, the switch unit 42 is connected to the terminal b to perform the video coding process. Further, in the case of interframe coding, ON / OFF control of the in-loop filter 50 is performed by selecting one of the terminals a and b of the switch unit 52.

The digitized image signal VI is input to the motion vector detection section 53, the subtraction circuit 51, and the switch section 42. The motion vector detection unit 53 extracts the motion vector information V required for motion compensation predictive coding. Further, in the subtraction circuit 51, the prediction signal VP is converted from the image signal VI.
Is subtracted to generate a prediction error signal in the case of interframe coding. The output signal PE of the switch unit 42 is subjected to a matrix calculation with a DCT conversion matrix (8 pixels × 8 lines) in the DCT calculation unit 43 to calculate a DCT coefficient. Then, the quantization unit 44 performs quantization with the characteristics of the quantization scale controlled by the coding control signal CT. Then, the entropy coding unit 45 performs variable coding processing in which a code having a short code length is assigned to a code having a high occurrence probability, and the coded data VE is processed.
To generate. On the other hand, the inverse quantizer 46 and the inverse DCT calculator 4
In addition, the prediction error signal PE ′ reproduced by performing the inverse quantization process and the DCT transform inverse matrix calculation in FIG.
The prediction signal VP is added at to reproduce the image signal. Then, the frame memory 49 uses the motion vector information V to perform motion compensation signal processing to generate a prediction signal.

FIG. 14 shows an embodiment of the information source decoding unit 13. When the coding type is intra-frame coding, the switch unit 57 receives the terminal a, according to the decoding control signal CTD.
In the case of interframe coding, the signal processing for video decoding is performed by connecting to the terminal b. The entropy decoding unit 54 converts the encoded data VE into the original fixed-length code. The inverse quantization unit 55 performs the inverse quantization process with the characteristics of the quantization scale determined by the signal CTD and decodes the DCT coefficient.
The inverse DCT calculator 56 calculates the DCT coefficient and the inverse DCT matrix and decodes the prediction error signal PE. Then, the adder circuit 60 adds it to the prediction signal VP to decode the original digitized image signal VO. Further, the prediction value generation unit 59 uses the motion vector information of the signal CTD to perform signal processing for motion compensation on the output signal of the frame memory 58 to generate a prediction signal.

CCITT Recommendation H. In the H.261 video coding, similarly to MPEG, a coded bitstream is hierarchically constructed, and in the header part of each layer, for example, a coding type, a quantization scale, a motion vector, a coding pattern And the like are assigned as the encoded parameter data CP. Then, the user data of the present invention generated by this encoded parameter data is inserted into the user data area. Then, the encoded data is added to form the encoded bitstream data. The user data generation unit of the encoding unit and the user data decoding unit of the decoding unit are the same as those shown in FIGS. 8 to 11 shown in the previous embodiment.
Since it can be realized by the configuration of, the description thereof will be omitted.

According to the present embodiment, CCITT Recommendation H.264 has little deterioration in image quality due to code errors occurring in the transmission / accumulation medium and high information efficiency compression efficiency. An image signal encoding / decoding device compliant with H.261 video encoding can be realized.

The present invention is based on the MPE described in the embodiment.
G, or H. The present invention is not limited to the video coding system compliant with H.261, but is generally applicable to a video signal coding / decoding device of a video coding system in which coding parameters and coded data are transmitted / stored. You can

[0037]

According to the present invention, by adding an extremely small amount of user data, it is less likely to be affected by a code error occurring in a transmission / storage medium, and the information amount of image data is several tenths. It is possible to realize an encoding / decoding device for an image signal that can be compressed as efficiently as possible.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention.

FIG. 2 is an explanatory diagram of MPEG video encoding.

FIG. 3 is an explanatory diagram of a coded bitstream of MPEG video coding.

FIG. 4 is an explanatory diagram of codes of a header section of an encoded bitstream.

FIG. 5 is an explanatory diagram of header part user data in the present invention.

FIG. 6 is an explanatory diagram of header part user data by shuffling processing.

FIG. 7 is a block diagram of a first embodiment of an information source coding unit.

FIG. 8 is a block diagram of a first embodiment of a user data generation unit.

FIG. 9 is a block diagram of a first embodiment of a user data decoding unit.

FIG. 10 is a block diagram of a second embodiment of a user data generation unit.

FIG. 11 is a block diagram of a second embodiment of a user data decoding unit.

FIG. 12 is a block diagram of a first embodiment of an information source decoding unit.

FIG. 13 is a block diagram of a second embodiment of an information source coding unit.

FIG. 14 is a block diagram of a second embodiment of an information source decoding unit.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Information source coding part, 2 ... Encoding control part, 2 ... User data generation part, 4 ... Video multiplex part, 5 ... Transmission buffer part, 6 ... Transmission path coding part, 7 ... Transmission / storage medium, 8 ... Transmission path decoding unit, 9 ... Reception buffer unit, 10 ...
Video multiplex unit, 11 ... User data decoding unit, 12 ... Decoding control unit, 13 ... Information decoding unit.

Claims (6)

[Claims] 1. An image signal encoding / decoding apparatus for transmitting / accumulating a digitized image signal by compressing an information amount by signal processing of high-efficiency encoding so that a high-efficiency encoding is performed by an encoding unit. A means for generating user data from encoded parameter data for signal processing and a means for adding the user data to an encoded bit stream are provided and transmitted to the decoding unit.
Coding / decoding of an image signal, characterized by comprising means for setting a coding parameter necessary for signal processing of decoding based on the coding parameter data of the coded bit stream and the user data accumulated Device. 2. The encoding of an image signal according to claim 1, wherein the user data generated from the encoding parameter is configured in a form capable of error detection or error correction for a code error occurring in the process of transmission / accumulation.・ Reproduction device. 3. A data structure of a coded bit stream according to claim 1 or 2, comprising a header part and a coded data part, and coding of an image signal to which coding parameter data and user data are allocated to the header part. -Decoding device. 4. The shuffling process according to claim 3, wherein a plurality of pieces of user data in the header portion within a specific period of the encoded bit stream are shuffled, and the encoding parameter data and the user data are different in each header portion. Device for encoding / decoding an image signal composed of encoding parameter data of a block. 5. The signal processing of high efficiency coding according to claim 1, 2, 3 or 4, is ISO / IEC JTC1 / S.
Standard MP for C29 / WG11 video coding for storage media
An image signal encoding / decoding device that is video encoding conforming to EG. 6. The high-efficiency coding signal processing according to claim 1, 2, 3 or 4, wherein CCITT Recommendation H.264 is used. An image signal encoding / decoding device that is video encoding conforming to H.261.