JP3159327B2 - Video data decoding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[Table of Contents] The present invention will be described in the following order. BACKGROUND OF THE INVENTION Problems to be Solved by the Invention Means for Solving the Problems (FIG. 3) Operation (FIGS. 4 and 5) Embodiment (FIGS. 1 to 6) Effects of the Invention

[0002]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving picture data decoding apparatus, and more particularly to a so-called compact disk or so-called DAT.
(Digital audio taperecorder) This is suitable for application to the case of reproducing moving image data from a cassette or a recording medium such as a hard disk.

[0003]

2. Description of the Related Art Conventionally, since the amount of information is extremely large for digitally recording a moving image, a recording medium having a very high continuous transmission speed is required for recording / reproducing the moving image. For example, when digitally recording a video signal according to the NTSC system, recording / reproducing is performed using a so-called video disk having a large amount of recorded information.

However, in order to record moving image data having a generated information amount similar to that of a so-called video disk for a long time on a smaller recording medium (that is, a recording information amount is small), a video signal is encoded with high efficiency. A means for recording and efficiently decoding the read signal is indispensable.

In order to meet such demands, a high-efficiency encoding method for an image signal has been proposed.
There is a G (Moving Picture Experts Group) system.
In the MPEG system, first, a difference between images is calculated in order to reduce the redundancy in the time axis direction, and then a discrete cosine transform (DCT) is performed in order to reduce the redundancy in the space axis direction.
It has been made to be.

[0006]

By the way, such M
In the PEG system, a decoding device is provided at a stage prior to decoding. However, if an error that cannot be corrected by the error code correction capability of the decoding device occurs, a partial loss of an image occurs at that point, and Displaying an image can be unsightly.

Therefore, two methods have been proposed in the past to correct the partial loss of the image. One is an inter-frame interpolation method in which a past frame image is corrected by fitting it into a missing portion as a partial image located at the same position as the missing image portion, and the other is interpolated from peripheral pixels of the missing image portion. This is an intra-frame interpolation method for correcting an image by fitting it into a missing portion.

However, the inter-frame interpolation method is effective in an image portion having little motion, but cannot smoothly connect a missing portion and a peripheral portion in a portion having a lot of motion.
There was a problem that good results were not always obtained due to noticeable omissions.

In contrast, the intra-frame interpolation method can correct the omission in the image part having a large motion so as not to be conspicuous in contrast to the case of the intra-frame interpolation method. There was a problem that the correction was unnatural.

The present invention has been made in view of the above points, and even when a transmission error in which a code error cannot be corrected on the decoding side occurs, the image loss is conspicuous regardless of the magnitude of the motion of the lost image. An object of the present invention is to propose a moving image data decoding device capable of decoding an image so that the moving image data cannot be decoded.

[0011]

According to the present invention, in order to solve the above-mentioned problems, moving image data S2 which is sequentially input is provided.
In the moving picture data decoding apparatus 20 for decoding the data 1, a code error of the moving picture data S 21 is corrected and a reproduced digital signal S 22 is output.
3; an error detecting / correcting means 23 for outputting 3; a demultiplexer means 24 for separating the motion vector data S30 and the image data S24 from the reproduced digital signal S22; a frame memory 29 for storing the frame data; Frame data generating means 28 for adding the motion compensation data S28 read from the frame memory 29 to the image data subjected to inverse quantization and inverse discrete cosine transform to generate decoded image data S29, and reference frame data read from the frame memory S32 and motion vector S
30, a motion compensating means 34 for generating predicted image data S33 from S31, a switching means 30 for stopping writing of the decoded image data S29 to the frame memory 29 when the error detection signal S23 is output, and a vector data S3.
0, and a switching means 3 for stopping the writing of the motion vector S30 to the motion vector memory 32 when the error detection signal S23 is output.
1, when the error detection signal S23 is output, the switching means 33 which inputs the motion vector motion compensation means 34 from the motion vector memory 32, and the decoded image data S42 read from the frame memory 29 and the delayed error Interpolated image data S43 is generated based on the detection signal S23 and the motion vector S41 read from the delayed motion vector memory 32, and the delayed motion vector data S4
It is determined whether the motion amount of the motion vector 1 is larger than a predetermined value. An interpolated image generating means for generating an interpolated image and, when the amount of motion of the delayed motion vector data S41 is smaller than a predetermined value, replacing the missing pixel data with a predicted image to generate an interpolated image; If the error detection signal S23 is not output, the decoded image data S generated by the frame data generation means 28 is output.
29 is output from the frame memory 29 via the switching means 30, and when the error detection signal S23 is output, the data S43 generated by the interpolation image generating means 40 is output from the frame memory 29. .

[0012]

When decoding a moving image data S21, if a code error that cannot be corrected is detected, an error detection signal S
23 is output by synthesizing the interpolated image data S43 or the predicted image data S33 with the image data S24 based on the motion amount of the motion vector S41 corresponding to the output video data S21.
Irrespective of the magnitude of the motion, the loss due to the code error can be corrected so as not to be conspicuous.

[0013]

BRIEF DESCRIPTION OF THE DRAWINGS FIG.

In FIG. 1, reference numeral 1 denotes a moving picture data encoding device (encoder) as a whole, and input picture data S1 obtained by converting an analog moving picture signal into digital data.
From the input terminal 2. At this time, the input image data S1 is composed of an intra frame I, a predictive frame P, and a bidirectional frame B as shown in FIG.

Here, the intra frames I (I ₀ , I ₁ ...)
..) Are frames which are compressed and transmitted only within the frame, and are predictive frames P (P ₀ , P _1).
……) are frames whose motion is predicted from one direction,
Bidirectional frame B (B ₀ , B ₁ , B ₂ ,
B ₃ ...) Are frames whose motion is predicted from both directions.

The difference data generation circuit 3 receives the input image data S1 from the input terminal 2 and
Further, the previous frame image data S2 of the previous frame stored in the frame memory 4 is input. Here, the difference data generation circuit 3 uses the input image data S
1 and the difference between the previous frame image data S2 and difference data S3 are generated, and the discrete cosine transform DCT
(Discrete cosine transform) Output to the circuit 5.

The discrete cosine transform circuit 5 performs discrete cosine transform of the difference data S3 in units of minute blocks using the two-dimensional correlation of the image, and transforms the resulting transformed data S4 to a quantization circuit Q (quantizer). 6
Output. The quantization circuit 6 quantizes the transformed data S4 with a predetermined quantization step size, and outputs the resulting quantized data S5 at an output terminal to a variable length coded VLC (variable length code) circuit 7.

The variable length coding circuit 7 outputs to the multiplexer 8 variable length coded data S6 obtained by subjecting the quantized data S5 to variable length coding. Multiplexer 8
Is the motion vector data S input from the encoder 9
7 is multiplexed with the variable-length coded data S6, and transmitted as transmission data S8 via the buffer circuit 10.

The moving picture data encoding device 1 has a local decoding circuit system 11 for locally decoding the quantized data S5 transmitted as the transmission data S8 and supplying it to the frame memory 4. I have. Local decoding circuit system 11
When the quantized data S5 is input to an inverse quantization circuit (Q ^-1 ) 12, the quantized data S5 is inversely quantized to a representative value, converted to inverse quantized data S10, and decoded data before quantization is decoded Then, the signal is supplied to a discrete cosine inverse transform (DCT ^-1 ) circuit 13.

Discrete cosine inverse conversion circuit 13
Is the inverse quantized data S1 decoded by the inverse quantization circuit 12.
0 is converted to decoded image data S11 by a conversion process reverse to that of the discrete cosine conversion circuit 5 and output to the frame data generation circuit 14. Here, the frame data generation circuit 14 restores the image data output as the transmission data S8 by adding the frame image data S2 fed back from the frame memory 4 and the decoded image data S11, and sequentially stores them in the frame memory 4. Has been made.

Further, when the moving picture data encoding apparatus 1 inputs the input picture data S1 to the motion vector calculating circuit 18 and obtains a motion vector, it supplies the motion vector S15 to the motion compensating circuit 19 and the encoder 9. ing. Here, the motion compensation circuit 19 includes the frame memory 4
Read out the decoded image data S16 from the memory and read out the motion prediction data S17 for motion-compensating the decoded image data S16.
Is output to the frame memory 4.

The encoder 9 encodes the motion data S15 obtained by the motion vector calculation circuit 18 and outputs the encoded data to the multiplexer 8 as motion vector data S7.

On the other hand, in FIG. 3, reference numeral 20 denotes a moving picture data decoding device (decoder) as a whole, and a reproduced signal S21 read from a recording medium is subjected to code error detection via an input terminal 21 via a buffer circuit 22. / Correction circuit 2
3 is input.

Here, the code error detection / correction circuit 23 detects and corrects errors contained in the image data from the reproduced data S21, and outputs the corrected reproduced image data S22 to the demultiplexer circuit 24. In addition, code error detection /
The correction circuit 23 outputs a switching signal S23 for controlling the switching of the output image data when detecting a block that cannot correct the code error.

The demultiplexer circuit 24 separates the motion vector data from the reproduced image data S22 and generates the difference vector information data S24 as a variable length decoding circuit (VLC ^-1 ).
25, and decodes the decoded image data S25 before being encoded by the variable-length code encoding circuit 7 to perform inverse quantization (Q
^-1 ) Supply to 26.

The inverse quantization circuit 26 outputs the decoded image data S2
5 is converted to inverse quantized data S26 by inverse quantization to a representative value. Discrete cosine inverse conversion circuit 27 converts the image data to decoded image data S27 by inverse conversion processing to discrete cosine conversion circuit 5 to generate frame data. Circuit 2
8 is output.

The frame data generation circuit 28 is provided with a motion compensation data S 28 read from the frame memory 29.
To the decoded image data S2
9 through the switching circuit 30 and the frame memory 29
It has been made to output more.

Here, the switching circuit 30 is turned on / off by a switching signal S23 supplied from the code error detection and correction circuit 23, and the reproduction data S21 is controlled by the code error detection and correction circuit 23.
When a code error that cannot be corrected is detected, the decoded image data S29 is not stored in the frame memory 29.

When the demultiplexer circuit 24 separates the vector data from the reproduced image data S22, it supplies it to the motion vector memory 32 via the switching circuit 31 as the current motion vector data S30.

Here, when the code error detection / correction circuit 23 detects no uncorrectable code error in the reproduction data S21, the switching circuit 31 stores the current motion vector data in the motion vector memory 32 holding the motion vectors of the past three frames. S30 is supplied.

The switching circuit 33 is controlled to be switched by a switching signal S23 supplied from the code error detection / correction circuit 23.
If no uncorrectable code error is detected, the current motion vector data S30 is supplied to the motion compensation circuit 34,
Alternatively, when an uncorrectable code error is detected, the previous motion vector data S31 stored in the motion vector memory 32 is supplied to the motion compensation circuit 34.

The motion compensation circuit 34 includes a frame memory 29
When the reference frame data S32 is input from the CPU, a predicted block image is generated based on the current or previous motion vector data S30 or S31, and is stored in the frame memory 29 as predicted image data S33.

When an uncorrectable code error occurs, the frame memory 29 temporarily replaces and stores the predicted image data S33 generated by the previous motion vector S31 as the corresponding block image data. , Output image data S34.

Here, when the motion amount of the previous motion vector data S31 is larger than a predetermined value, the frame memory 29 replaces the interpolation image data S43 newly input from the interpolation circuit 40 with the prediction image data S33. Remember,
The output is performed as output image data S34.

The interpolation circuit 40 receives the motion vector data S41 from the motion vector memory 32 via the delay circuit 41, and receives the switching signal S23 via the delay circuit 42.
Is input, and if the motion amount of the motion vector data S41 is larger than a preset threshold value, the image data S42 of four pixels read from the frame memory 29
Interpolated image data S43 is generated on the basis of (pixels A, B, C, and D of adjacent blocks when a pixel having a loss in FIG. 4 is X).

In this case, the interpolation circuit 40 performs image processing (a, b, c, d) for four pixels located in four blocks above, below, left, and right of the block R of interest in order to perform the intra-frame interpolation processing.
Therefore, the decoding of the block adjacent to the lower side of the target pixel R is waited for, and the pixel data is interpolated. Incidentally, when the block of interest R is located at the bottom row of the screen, the interpolation circuit 40 immediately starts the intra-frame interpolation processing after the inter-frame interpolation processing by the motion compensation circuit 34 is completed.

Here, the delay circuits 41 and 42 provide the motion vector data S41 so that the interpolation circuit 40 can correct the block one column (one slice) past in the block where decoding has been completed or inter-frame correction has been completed. And the switching signal S23 is delayed.

In the above arrangement, the moving picture data decoding apparatus 20 sequentially reads out the reproduced data S21 from the recording medium and inputs the read data S21 to the code error detection and correction circuit 23 via the buffer circuit 22.

Here, if the code error of the reproduced data S21 can be corrected by the code error detection and correction circuit 23, the moving picture data decoding device 20 decodes the corrected image data S22 corrected based on the error detection and correction code. The signal is supplied to a frame data generation circuit 28 through a multiplexer circuit 24, a variable length decoding circuit 25, an inverse quantization circuit 26, and a discrete cosine inverse transformation circuit 27 in order.

The frame data generation circuit 28 performs motion compensation on the intra-frame I0 stored in the frame memory 29 as a reference frame by motion compensation data S2.
8 to the decoded image data S29.
Is decoded, the frame memory 2
9, and sequentially decodes the image data of the predictive frame P0 and the bidirectional frames B0, B1,... And outputs the decoded data as decoded image data S34.

On the other hand, the moving picture data decoding apparatus 20
When the block which cannot correct the code error of the reproduction data S21 is detected by the code error detection and correction circuit 23,
The switching signal S23 is output to the switching circuits 30 and 31, and writing of the decoded image data S29 and the current motion vector data S30 to the frame memory 29 and the motion vector memory 32 is stopped.

Further, the moving picture data decoding device 20 outputs the switching signal S23 to the motion vector memory 32 and the switching circuit 33.
The motion vector in the corresponding block of the past frame stored in the motion vector memory 32 is read out, supplied to the motion compensation circuit 34, and replaced with the corresponding pixel by the predicted image data S33 predicted from the past motion vector.

At this time, the interpolation circuit 40 determines whether or not the motion amount of the motion vector data S41 input via the delay circuit 41 is larger than a predetermined value. In the frame memory 29, the pixel data x of the missing pixel X is stored in the predicted image data S33.
Then, the interpolated frame image is output as output image data S34.

On the other hand, when the motion amount is larger than the predetermined value, the interpolation circuit 40 determines that the pixel X in which a code error that cannot be corrected is generated has four pixels (A) adjacent to the block R of interest in which the target block R has occurred. , B, C, D) to generate image data of the pixel of interest X and output it as interpolated image data S43.

Here, the four pixels A, B,
C and D are pixels located in the same column or the same row as the target pixel X and adjacent to the target block X, and have pixel values a, b, and
c and d. If the distances between the four pixels A, B, C, and D and the target pixel X are LA, LB, LC, and LD, the pixel value x of the target pixel X is expressed by the following equation.

(Equation 1) Is obtained by linear interpolation using interpolation image data S43
Is output to the frame memory 29.

At this time, the frame memory 29 stores the pixel value x of the target pixel X in which a code error that cannot be corrected has occurred in the interpolated image data S43 generated by the intra-frame interpolation in place of the temporarily stored predicted image data S33. It is newly rewritten and output as output image data S34.

On the other hand, when the block of interest R including the pixel of interest X in which a code error for which error correction cannot be performed has occurred is the outermost block of the screen, or one of adjacent blocks has a code error for which error correction cannot be performed. In the case of (4), since four pixels A, B, C, and D cannot be referred to from the adjacent block, the pixel value x is obtained by using the pixel of the adjacent block on the opposite side to the undetectable adjacent block.

For example, in FIG. 4, when the pixel value b of the pixel B cannot be obtained because the adjacent block on the left side cannot be referred to,

(Equation 2) Is substituted into the expression (1) to obtain the pixel value x of the target pixel X.

When it is not possible to refer to the adjacent blocks located on the left and right sides of the target block R including the target pixel X in which a code error in which error correction is not possible has occurred,

(Equation 3) To obtain the pixel value x, and corrects the image of the missing pixel so that it is not conspicuous in the peripheral portion.

According to the above arrangement, in a moving image data decoding apparatus for decoding and reproducing moving image data in block units, when a code error which cannot be corrected in the reproduced moving image data is detected, The corresponding block image is temporarily replaced with a predicted image generated from the past frame based on the motion vector in the past frame. By replacing the interpolated image with the interpolated image, the missing image can be corrected so that the loss due to the code error is not conspicuous irrespective of the magnitude of the motion of the image in which the loss occurs.

In the above-described embodiment, a case has been described where motion is predicted for an image in units of frames. However, the present invention is not limited to this, and can be applied to a system for performing motion prediction in units of fields.

Further, in the above-described embodiment, a case has been described where the present invention is applied to a system using discrete cosine transform DCT and motion prediction between frames. However, the present invention is not limited to this, and prediction processing between frames is performed. This is suitable for application to a moving image data decoding device that executes

Further, in the above-described embodiment, when interpolating image data having a code error in which error correction cannot be performed, interpolation is performed using image data motion-compensated by a past motion vector corresponding to the image data. As described above, the present invention is not limited to this, and pixel data in a corresponding block of a past frame may be replaced.

Further, in the above-described embodiment, when a code error that cannot be corrected occurs, a case has been described in which intra-frame or inter-frame interpolation processing is performed using a motion vector of a past frame. However, the present invention is not limited to this. Interpolation processing may be performed using motion vectors of adjacent blocks in the same frame. For example, the motion vector of the block on the left or upper side where decoding has been completed for the block of interest R may be used.

Further, in the above-described embodiment, when it is not possible to refer to the adjacent blocks located on both the left and right sides of the target block R including the target pixel X in which a code error that cannot be error-corrected has occurred, the formula (3) is used. The case where the pixel value x of X is obtained has been described. However, the present invention is not limited to this, and when it is not possible to refer to the adjacent block vertically adjacent to the target block R, the pixel value x of the target pixel X is similarly determined. You can ask.

Further, in the above-described embodiment, the sequence of the configuration shown in FIG. 2, that is, I ₀ , B ₀ , B ₁ ,
The case of P ₀ , B ₂ , B ₃ , P ₁ , B ₄ , B ₅ , I ₁ ... Has been described, but the present invention is not limited to this, and the intra-frame I, the pre-directory frame P, and the bi-directional frame The present invention can also be applied to sequences in various combinations using the nally frame B.

Further, in the above-described embodiment, the case where the motion vectors of the past three frames are stored in the motion vector memory 32 has been described. However, the present invention is not limited to this. Widely applicable.

[0058]

As described above, according to the present invention, when a code error that cannot be corrected is detected, the image is corrected according to the motion amount of the motion vector corresponding to the moving image data in which the code error is detected. Data and predicted image data (obtained based on reference frame data and motion vector) or image data and interpolated image data (obtained based on decoded image data, delayed error detection signal and delayed motion vector), and Is synthesized and decoded, the missing part of the image due to the code error can always be smoothly corrected with respect to the peripheral image according to the magnitude of the motion of the decoded image.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of a moving image data encoding device according to the present invention.

FIG. 2 is a schematic diagram for explaining the inter-frame motion prediction.

FIG. 3 is a block diagram showing a configuration of an embodiment of a moving picture data decoding apparatus according to the present invention.

FIG. 4 is a schematic diagram for explaining the block interpolation.

FIG. 5 is a schematic diagram for explaining an intra-frame interpolation process;

FIG. 6 is a schematic diagram for explaining another embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Moving image data encoding device, 20 ... Moving image data decoding device, 21 ... Input terminal, 22 ... Buffer circuit, 23 ... Code error detection / correction circuit, 24 ... Demultiplexer circuit, 25 …… Variable length decoding circuit, 26 ……
Inverse quantization circuit 27 Discrete cosine inverse transformation circuit 28 Frame data generation circuit 29 Frame memory 30, 31, 33 Switching circuit 32
Motion vector memory, 34 ... motion compensation circuit, 40 ...
Interpolation circuits, 41, 42... Delay circuits.

Claims (2)

(57) [Claims] A moving picture for decoding moving picture data inputted sequentially.
In the image data decoding device, a code error of the moving image data is corrected, and a reproduction digital
Outputs a signal and corrects the moving image data by error correction.
If an undetectable code error is detected, an error detection signal is output.
Detection / correction means, and motion vector data and image data from the reproduced digital signal.
Demultiplexer means for separating data from data;A frame memory for storing frame data; Variable length decoding, inverse quantization, inverse discrete cosine transform
The image data read from the frame memory
Frame to generate decoded image data by adding compensation data
Means for generating frame data; Reference frame data read from the frame memory
Compensation to Generate Predicted Image Data from Motion and Motion Vector
Means, When the error detection signal is output, the decoded image data is output.
To stop writing data to the frame memory
Means, A motion vector memory for storing the vector data, When the error detection signal is output, the motion vector
To stop writing to the above motion vector memory.
Exchange means, When the error detection signal is output, the motion vector
Switching from memory to motion vector motion compensation means
Means, The decoded image data read from the frame memory and
Delayed With the above error detection signalThe above motion vector is delayed
Motion vector read from theOn the basis of the
Generate interpolated image dataAnd the motion vector is delayed
Judge whether the amount of data movement is larger than the specified value and
When the motion amount of the motion vector is larger than a predetermined value,
Indicates the missing image read from the frame memory.
Generates an interpolated image based on four pixels adjacent to the element, and delays
The motion amount of the motion vector data thus obtained is smaller than a predetermined value.
In this case, the missing pixel data is
To generate an interpolation imageInterpolation image generation means,With  If the error detection signal is not output,Frey
The decoded image data generated by the
Output from the frame memory via the switching means
So thatWhen the error detection signal is output,Up
The data generated by the interpolation image generating means is
Output from frame memoryCharacterized by
Moving image data decoding device. 2. The moving picture data decoding apparatus according to claim 1, wherein said sequentially inputted moving picture data is moving picture data read from a disk-shaped optical recording medium.