Encoder and decoder for moving image

Description: TECHNICAL FIELD [Detailed description of the invention]

【0001】

[Industrial application field]

The present invention is a high-efficiency coding / decoding device that is used in recording, transmission, and display devices that process digital signals and efficiently encodes / decodes, and the image quality deteriorates even when a transmission code error occurs. The present invention relates to a coding device and a decoding device that perform inter-image processing of a small number of moving image signals.

【0002】

[Conventional technology]

In high-efficiency coding of a moving image signal (moving image), a frame that uses the correlation between frames of the image signal to predict the coded frame with the coded frame and encodes only the prediction error. There is inter-predictive coding. In recent years, motion compensation frame-to-frame prediction, which predicts by moving an image according to motion, has become more common. On the other hand, in coding for storage media, intra-frame independent coding is performed every few frames without inter-frame prediction to support random access and high-speed search. Further, a method in which the coding efficiency is improved by jump prediction and front-back prediction between them is known as an MPEG (ISO-IEC) method. In the MPEG method, each frame is an I (Intra) frame that is independently encoded within the frame, a P (Prediction) frame that is predicted to jump, and a B (Bi-directional) frame that is predicted from the front and back due to the difference in the prediction method. There is.

【0003】

<Conventional example (coding device)> Fig. 4 shows a configuration example of the MPEG type coding device. Here, depending on the frame types I, P, and B, the changeover switches 2, 4, and 22 are controlled by the synchronization signal separated from the input signal, and are switched to the one shown in the figure. The image signal input from the image input 1 is directly guided to the subtractor 5 through the changeover switches 2 and 4 in the I and P frames, and the I and P before and after the B frame are encoded in the frame memory 3. After being delayed to, it is guided to the predictive subtractor 5. In the prediction subtractor 5, the prediction signal coming from the adaptive predictor 42 is subtracted from the input signal to output the prediction residual signal, which becomes the coded data encoded and compressed by the in-frame encoder 6.

【0004】

In the in-frame encoder 6, DCT (discrete cosine transform) is first performed, the conversion output is quantized, and variable length coding is performed using a Huffman code or the like. The compressed DCT information is given to the multiplexer 40, and in the I and P frames, it is guided to the in-frame decoder 21 through the changeover switch 22. The in-frame decoder 21 first decodes the variable-length code, replaces the fixed-length code with a quantization representative value, and then reverse DCTs to obtain a reproduction signal. The predicted residual signal reproduced by the in-frame decoder 21 becomes an image signal reproduced by adding the predicted signal by the residual adder 20. The reproduced image signal is stored in the frame memory 19, and the signal previously stored in the frame memory 19 is transferred to the frame memory 18.

【0005】

The output of the frame memory 19 is given to the motion compensator 15 and the motion vector detector 17, and the output of the frame memory 18 is given to the motion compensator 14 and the motion vector detector 16. The motion vector detectors 16 and 17 detect a motion vector between the input signal and the signal stored in the frame memories 18 and 19 for each block of 16 × 16 pixels. The motion vector information is given to the motion compensators 14 and 15 as well as to the multiplexer 40. The motion compensators 14 and 15 spatially move the reproduced image signal stored in the frame memories 18 and 19 by the motion vector given by the motion detector and give it to the adaptive predictor 42.

【0006】

The adaptive predictor 42 creates four types of prediction signals from two motion-compensated signals (F and B) in the same block as motion vector detection, and matches the optimum one with the input signal that is the predicted signal. Judge from. Then, the optimum prediction signal is output, and information on which is selected (prediction mode) is given to the multiplexer 40. There are four prediction modes used here: F (prediction signal from the previous frame in time) only, B (prediction signal from the later frame in time) only, (F + B) / 2, 0. Yes, 0 results in unpredictable in-frame independent coding. Only 0 is used in the prediction mode, F and 0 are used in the P frame, and all four types are used in the B frame.

【0007】

The multiplexer 40 contains DCT information which is the output of the in-frame encoder 6, prediction mode information (MODE) which is the output of the adaptive predictor 42, and motion vector information (MVF and MVB) which are the outputs of the motion vector detectors 16 and 17. ) Is grouped for each block (macroblock: MB) for which the motion vector and the prediction mode are required, and is output to the decoding device side via the data output 12. The data structure is as shown in Fig. 6 (A). Here, the motion vector information is not transmitted for those that are not used in the prediction.

【0008】

<Conventional example (decoding device)> The configuration of the decoding device is shown in FIG. The same parts as those of the coding device shown in FIG. 4 are designated by the same reference numerals. The coded data input from the data input 30 is decomposed into each information by the demultiplexer 41, the DCT information is sent to the in-frame decoder 21, the prediction mode information is sent to the adaptive predictor 43, and the motion vector information is sent to the motion compensator 14. , 15 is given. The DCT information is decoded by the in-frame decoder 21 and the prediction signal is added by the residual adder 20 to become a reproduced image signal.

【0009】

This reproduced image signal is output immediately from the reproduced image signal output 36 via the changeover switches 34 and 35 in the B frame, and is stored in the frame memory 19 in the I and P. The signal previously stored in the frame memory 19 is transferred to the frame memory 18 and output from the reproduced image signal output 36 via the changeover switch 35.

【0010】

The output of the frame memory 19 is given to the motion compensator 15, and the output of the frame memory 18 is given to the motion compensator 14. The motion compensators 14 and 15 spatially move the reproduced image signal stored in the frame memory by the motion vector given by the demultiplexer 41 and give it to the adaptive predictor 43. The adaptive predictor 43 creates a prediction signal from the prediction mode information given by the demultiplexer 41 and outputs it to the residual adder 20. Here, the inter-image processing unit is a frame, but the same applies to the interlaced signal field.

【0011】

[Problems to be Solved by the Invention]

If a code error occurs in transmission or recording between the coding device and its decoding device, normal decoding will not occur and image quality will deteriorate. Code errors include those that occur on ordinary lines and recording media and cell losses on ATM lines, and the latter is missing on a cell-by-cell basis. In this case, even with the coding device and its decoding device shown as a conventional example, it is usually detected where the code error has occurred. Therefore, in the case of an error in the predicted residual, the predicted residual is not added and the predicted signal is used as it is as the reproduced image. For example, the deterioration is not so great. However, if the motion vector is missing, the block cannot be decoded at all, and interpolation is performed from adjacent blocks in the same frame, which inevitably deteriorates the image quality.

【0012】

On the other hand, the coding method having an independent frame periodically seems to be advantageous at first glance because the deterioration up to that point stops at the independent frame. However, the code error of the independent frame can be interpolated only spatially, resulting in a large deterioration and affecting the subsequent images. Furthermore, since the amount of data in the independent frame is larger than that in the predicted frame, even if the independent frame is once in 10 frames, the amount of data occupies about 40% of the total, and the influence of the code error is rather serious.

【0013】

The present invention is intended to solve such a problem, and the frames used for the inter-image processing are always made into a plurality of frames, and the information of the motion compensation and the inter-image processing method is collected for each type of information from the encoding device. To transmit. The decoding device detects an error and switches to another frame without using the incorrect frame in the inter-image processing for each block. As a result, it is an object of the present invention to provide a moving image coding device and a decoding device that do not cause a large deterioration in image quality even if a code error occurs in a transmission line.

【0014】

[Means for solving problems]

In order to achieve the above object, in decoding a moving image signal, it is a moving image signal decoding device encoded by dividing one frame or a field into a plurality of blocks, and processing a transmission code error between images. Error detection means 38 that detects the type of code indicating the method, the type of code of each motion vector, and the type of code for in-image processing to identify the type of code with an error, and between frames or fields for each block. When performing motion compensation inter-image processing between images, the inter-image processing means 33 that performs decoding processing by changing the inter-image processing method to one that uses only the error-free code according to the information on the type of the code having the error. A moving image decoding apparatus having the above and 37 elements is provided.

【0015】

Further, it is a coding device that divides one frame or a field into a plurality of blocks and encodes the moving image signal, and performs motion compensation inter-image coding processing between frames or between fields for each block. Then, a code dedicated to error countermeasures indicating the motion compensation and the inter-image processing method used only when a transmission code error occurs and a normal code indicating the motion compensation and the inter-image processing method used when the transmission code error does not occur are used. A moving image coding device having inter-image processing means 5, 6, 13 to 17 to be output, and a multiplexing means 11 for time-dividing and multiplexing the code dedicated to error countermeasures and the normal code and outputting them in one system. Also, an error countermeasure indicating a motion compensation and inter-image processing method used only when a transmission code error occurs, which is a video signal decoding device transmitted from the video coding device described above. Means 7 to 10 for receiving a dedicated code and a normal code indicating a motion compensation and image-to-image processing method used when the transmission code error does not occur, and a block having a code error by detecting a transmission code error for each block. When the error detecting means 38 for specifying the error detection means 38 and the block having the code error specified by the error detecting means when the motion compensation image-to-image decoding process is performed between frames or fields for each block, the code dedicated to the error countermeasure is used. Provided is a moving image decoding apparatus having inter-image processing means 33 and 37 for performing decoding processing using the above.

【0016】

[Action]

In the coding device, the frames used for the inter-image processing are always made into a plurality of frames, and the information of the motion compensation and the inter-image processing method is collectively transmitted for each type of information, so that the error detection is performed in a fixed amount of code units. Even if it is done, the probability that multiple pieces of information cannot be used in the same block is low.

【0017】

The decoding device detects transmission code errors for each type of coded information and switches to another frame without using the signal of the frame with the incorrect information in the inter-image processing for each block, so the deterioration of image quality is extremely high. Only a few.

【0018】

[Example]

An embodiment of the moving image coding device and the decoding device thereof according to the present invention will be described in detail below with reference to the drawings. <First Example (Encoding Device)> FIG. 1 is a block diagram showing a first embodiment of the moving image coding device of the present invention. The same parts as those of the conventional example described in detail in FIG. 4 are designated by the same reference numerals. In FIG. 1, the basic coding process is the same as in the conventional example, and the changeover switches 2, 4, 22, the predictive subtractor 5, the in-frame encoder 6, the in-frame decoder 21, the residual adder 20, The operations of the frame memories 18, 19, the motion compensators 14, 15, and the motion vector detectors 16, 17 (points of encoding in block units, etc.) are the same as in the conventional example.

【0019】

This coding device has DCT information which is the output of the in-frame encoder 6, prediction mode information (MODE) which is the output of the adaptive predictor 42, and motion vector information (MVF) which is the output of the motion vector detectors 16 and 17. And the transmission method of each information such as MVB) is different from the conventional example (Fig. 4). That is, the memories 7, 8, 9, 10 and the selector 11 are configured so that each information is collected in the memory and then transmitted.

【0020】

The DCT information that is the output of the in-frame encoder 6, the prediction mode information that is the output of the adaptive predictor 13, and the motion vector information that is the output of the motion vector detectors 16 and 17, are stored in memories 7,8,9,10, respectively. It is stored once. Then, when 30 to 300 macroblocks are collected, they are selected by the selector 11 and sequentially output via the data output 12 in the form shown in (B) of FIG. The size of collecting the coding information in units of a plurality of blocks (macroblocks) for each type may be set sufficiently larger than the error detection block and smaller than the number of blocks in one frame. In this embodiment, the missing cell unit on the ATM line is for several macroblocks, and the number of blocks for one frame is about 1350 macroblocks. Further, in the adaptive predictor 13, in the conventional example, only one frame (F frame) was used in the P frame, but in this embodiment, a plurality of frames (both F and B) are used. ing. In order to deal with errors in the decoding device, the motion vector is sent as a code dedicated to error countermeasures even if it is not used in normal decoding in the prediction mode.

【0021】

<First Example (Decoding Device)> FIG. 2 is a block diagram showing an example of a decoding device corresponding to the coding device of FIG. 1 of the present invention. The same parts as those in FIGS. 1 and 5 are indicated by the same reference numerals. In FIG. 2, the basic coding process is the same as that of the conventional example of FIG. 5, and the operations of the changeover switches 34,35, the in-frame decoder 21, the frame memory 18,19, and the motion compensator 14,15 are the conventional examples. Is the same as.

【0022】

The difference from the conventional example (Fig. 5) is the memory 7,8,9,10, selector 31, error detector 38, and variable adder 33 that adds the in-frame decoding signal and the signal from the predictor with different gains. In a certain point, the handling of each information and the operation of the adaptive predictor 37 and the variable adder 33 are different. Be other words, through the data input terminal 30 and incoming transmitted coded data signal from the encoding apparatus, DCT information has been multiplexed by the selector 31, the prediction mode information, motion vector information are separated, respectively It is stored once in memory 7,8,9,10. Then, the DCT information stored in the memory 7 is stored in the in-frame decoder 21, the prediction mode information stored in the memory 8 is stored in the adaptive predictor 37, and the motion vector information stored in the memories 9 and 10 is moved. Given to compensators 14,15.

【0023】

On the other hand, the error detector 38 determines which type of information in which macroblock is incorrect based on the decoding of the error detection code of the transmission line coding and the cell loss information on the ATM line. The prediction mode is changed according to the rules in Table 1 depending on the error. In Table 1, C is the current decoding signal that is the output of the in-frame decoder. If the mode information is missing, the frame of B or F that is closer to C in time is used.

【0024】

[table 1]

<img file="JP2962012B2_D0001.tif" />【0025】

If there is a normal code for a frame of error-free information in a macroblock that uses multiple images as shown in Table 1, discard the incorrect ones and make only the correct ones. If the normal code originally used in the decoding process has an error for all frames, the code dedicated to error countermeasures for the frame originally not used in the decoding process is also used. The operation of the variable adder 33 is usually the same as that of the residual adder 20 of the conventional example. When the DCT information is missing, the output of the in-frame decoder 21 is set to 0 and the output of the adaptive predictor 37 is output as it is. To do.

【0026】

In this way, according to the main image coding device and its decoding device, a plurality of frames used for inter-image processing are always used, and information on their motion compensation and inter-image processing method is collectively transmitted for each type of information. However, the decoding device detects errors for each type of information and switches to another frame without using the signal of the frame with the incorrect information in the inter-image processing for each block, so error detection is a fixed amount of code. Even if it is performed in units of, the probability that multiple pieces of information cannot be used in the same block will be low, and the deterioration of image quality will be extremely small due to the limitation or switching of the frames used.

【0027】

&lt;Second Example (Decoding Device)&gt; The second embodiment of the present invention is a prior application (Japanese Patent Application No. 3-125393) by the same inventor and the same applicant as the present invention. The present invention is applied to the apparatus disclosed in "Chemical Equipment", and a coding error is dealt with by changing the inter-image processing even in an independent frame. This improves the coding efficiency while maintaining the independence of the frames. Even in the I frame, the inter-frame correlation is used, and the inter-frame image addition is performed by the decoding device to maintain the image quality even if the quantization is roughened. It is based on a coding method that can be used. The decoding device differs from the first embodiment, and its configuration is shown in FIG. The input data is encoded by the coding device shown in FIG. The decoding device of FIG. 3 differs from that of FIG. 2 in that it has a matching determination device 32 that determines matching between two images. That is, both the output of the in-frame decoder 21 and the output of the adaptive predictor 36 are configured to be guided by the matching determiner 32 and the variable adder 33.

【0028】

Further, the operation differs from that of the first embodiment only in the I frame, and the P and B frames are the same. In the I frame, the matching determiner 32 examines the matching of the two images and gives information to that extent to the variable adder 33. In the variable adder 33, when the matching is good, the prediction signal from the adaptive predictor 37 is increased, and when the matching is bad, the current frame signal from the in-frame decoder 21 is increased and added. Here, the sum of the gains of each signal is 1. If there is an error in the signal from the in-frame decoder, the output of the in-frame decoder 21 is forcibly set to 0, and only the output of the adaptive predictor 37 is used as the reproduction signal. As a result, the code error that occurs in the independent frame is also covered. As a coding device corresponding to the decoding device of the second embodiment, compensation for the quantization error of the I frame may be used.

【0029】

[Effect of the invention]

In the moving image coding device and its decoding device of the present invention, a plurality of frames used for inter-image processing are always used, and information on their motion compensation and inter-image processing method is collectively transmitted for each type of information, and the decoding device is used. Then, by detecting an error for each type of information and switching to another frame without using the signal of the frame with the incorrect information in the inter-image processing for each block, the error detection is performed in a certain amount of code units. Even so, the probability that multiple pieces of information cannot be used in the same block is low, and the deterioration of image quality is extremely small due to the limitation or switching of the frames used. As a result, even if a considerable code error occurs in the transmission line, the image quality does not deteriorate significantly. Therefore, since code errors are allowed, transmission line coding does not require a large amount of error correction code, and as a result, the amount of data can be reduced. As described above, the moving image coding device and the decoding device of the present invention have extremely excellent effects in practical use.

[Simple explanation of drawings]

[Figure 1]

It is a block diagram which shows 1st Example of the moving image coding apparatus which becomes this invention.

[Figure 2]

It is a block diagram which shows 1st Example of the moving image decoding apparatus which becomes this invention.

[Fig. 3]

It is a block diagram which shows the 2nd Example of the moving image decoding apparatus which becomes this invention.

[Fig. 4]

It is a block diagram which shows the conventional example of a moving image coding apparatus.

[Fig. 5]

It is a block diagram which shows the conventional example of the moving image decoding apparatus.

[Fig. 6]

It is a figure which shows the data structure of the prior art example and the Example.

[Explanation of symbols]

1 ... image signal input, 2, 4, 22, 34, 35 ... selector switch, 3, 18, 19 ... frame memory, 5 ... predictive subtractor, 6 ... in-frame coding Instrument, 7, 8, 9, 10 ... memory, 11, 31 ... selector, 12 ... data output terminal, 13, 37, 42, 43 ... adaptive predictor, 14, 15 ... Motion Compensator, 16, 17 ... Motion Vector Detector, 20 ... Residual Adder, 21 ... In-frame Decoder, 30 ... Data Input Terminal, 32 ... Matching Detector, 33 ... variable adder, 36 ... playback image signal output, 38 ... error detector, 40 ... multiplexer, 41 ... demultiplexer