JP3257226B2 - Image processing apparatus and image processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and an image processing method using block coding, and more particularly to a process for extracting variable-length data when there is an error in decoding data.

[0002]

2. Description of the Related Art When compressing a data transmission amount of digital image data, a two-dimensional cosine transform (Discrete Cosin
An encoding method using an orthogonal transform such as e Transform (hereinafter, referred to as DCT) has been conventionally proposed.

[0003] In the coding method using DCT, a television signal of one frame is divided into a plurality of small blocks each consisting of n pixels in the horizontal direction x m pixels in the vertical direction, and DCT is applied to each block. The obtained DC component coefficient data and a plurality of AC component coefficient data are converted into entropy codes having different bit lengths, for example, Huffman codes, according to the appearance probabilities of the AC components, and transmitted.

[0004] By the way, it is conceivable to record the image data compressed as described above on a digital VTR. In this digital VTR, data is recorded on a magnetic tape. Errors often occur in data due to scratches and the like. For this reason, digital VTRs are generally
Performs error correction using a Reed-Solomon code or the like. This error correction greatly reduces the number of errors. However, if there are many original errors, the error cannot be restored to its original state by error correction alone, and errors remain in the data. Generally, data that cannot be restored by error correction is set with an error flag in units of bytes to distinguish the data from data having no error.

In the case of a digital VTR that does not use compression, concealment is performed on erroneous data, for example, by interpolating using surrounding data by utilizing image correlation. For example, the average value of the four pixel data at the top, bottom, left and right of the error data is replaced with this error data.

[0006]

SUMMARY OF THE INVENTION However, DCT
In the encoding method according to the above, since data is continuously recorded in a Huffman code with a variable bit length, if a bit error occurs, correct code identification cannot be performed, and subsequent data extraction also fails. Very likely. At that time, the data after the code in which the error occurred is basically all incorrect data.

Also, as a feature of the transform coding, if any one of the components in a block has an error, the effect is not limited to the pixel unit but extends to the entire block. For this reason, when concealing, instead of targeting pixels,
Must be performed on blocks. However, it is difficult to conceal the entire block, and the deterioration is very conspicuous.

In order to solve the above-mentioned problem of the so-called propagation error, there has been proposed a method of estimating missing coefficients from error-free coefficient data which can be picked up. In this method, the missing coefficient of the target block is estimated using the error-free coefficient data of the target block and the data of the peripheral blocks of the target block.

However, to realize this method,
Some of the key coefficients in the block need to be picked up without errors. In order to guarantee this, it is necessary to use a fixed length record without using the Huffman code for the main coefficient in order to avoid data from being unable to be cut out. The main coefficient means coefficient data having a large influence on the decoded image of the block. As a result, there is a disadvantage that the coding efficiency is deteriorated. In addition, a constraint condition that there is no error in the data of the peripheral block of the target block is required.

Further, as a fundamental problem, in the above-mentioned method, since other coefficients are estimated from a few obtained coefficients at most, the accuracy of the obtained coefficients is extremely poor. However, there is a disadvantage that a good restored image cannot be obtained.

Accordingly, an object of the present invention is to provide an image processing apparatus and an image processing method which solve the above problems.

[0012]

According to a first aspect of the present invention, there is provided an image processing apparatus for processing coded image data which has been subjected to variable length coding for each block . Error flag obtaining means for obtaining an error flag indicating presence / absence, and all possible values of encoded image data of a predetermined unit indicated by the error flag.
It gives the data value of Te, the given encoded image data of each data value, and decoding means for decoding in accordance with the terms of the variable length coding, restores data of each data value decoded by the decoding means
Includes a motor, and a selection means for correlation with restoration data adjacency block selects the data value as a maximum, correlation, block
An image processing apparatus characterized in that the image data is obtained between restored data located in the vicinity of a data boundary . An image processing method according to claim 5, wherein in the image processing method for processing coded image data that has been subjected to variable-length coding for each block, a step of acquiring an error flag indicating presence or absence of an error for each predetermined unit of the coded image data. And all possible data values for the coded image data in the predetermined unit indicated by the error flag are provided, and the coded image data given each data value is converted to a variable length coding condition. decoding step and restoring data of each decoded data values in the decoding step of decoding according to
It includes a motor, a step of correlation of the restored data of the neighbor block is to select the data value as a maximum, correlation, block
This is an image processing method characterized in that it is obtained between restored data located in the vicinity of a boundary of a mark .

Further, the present invention is intended to determine not only the correlation coefficient of the target block but also the correlation coefficient of the restored data of the adjacent block.

[0014]

Originally, an image has a very strong correlation between adjacent pixels. Even when an image has been coded to a high degree of efficiency, a strong correlation is maintained when the image is decoded. However,
In blocks decoded from erroneous coefficients, the correlation is reduced. In particular, when there is an error in the coefficient data of the low-frequency AC component, the tendency is remarkable. If there is an error in one byte of the variable length data, the original value of the data of that byte is not certain, but it is certain to take any value from 0 to 255. Therefore, the data is changed in 256 ways from 0 to 255, the data is cut out and decoded, the pixel correlation in the block is obtained, and the one having the maximum pixel correlation is selected. It is likely to be right or very close to the right one. Therefore, by outputting data that maximizes the pixel correlation, it is possible to output a decoded value that is very likely to be close to the correct one.

In addition, when the correlation is determined, the error in the determination can be prevented by including the restored data of the adjacent block.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which a high-efficiency encoding method according to the present invention is applied to a digital VTR will be described below with reference to the drawings. The configuration on the reproduction side of the digital VTR will be described with reference to FIG. In FIG. 1, reproduction data from a magnetic head 1 is transmitted to a channel decoder 3 via a rotary transformer (not shown) and a reproduction amplifier 2.
Respectively. In the channel decoder 3, the channel coding is demodulated, and the output signal of the channel decoder 3 is converted to a TBC circuit (time axis correction circuit).
4 is supplied. In this TBC circuit 4, the time axis fluctuation component of the reproduction signal is removed.

The reproduced data from the TBC circuit 4 is supplied to an ECC circuit 5, where error correction using an error correction code is performed. The output signal generated from the ECC circuit 5 includes an error flag indicating the presence or absence of an error after error correction, in addition to the reproduced data. An output signal of the ECC circuit 5 is supplied to a frame decomposition circuit 6. The frame decomposition circuit 6 separates each component of the block coded data of the image data, and switches the recording system clock to the image system clock. The present invention relates to this frame decomposition circuit.

The output signal of the ECC circuit 5 is composed of a plurality of sync blocks SB. FIG. 2 shows an example of the configuration of the sync block SB. The sync block SB is formed from a plurality of transmission unit blocks BL. This transmission unit block BL has a fixed length, and is composed of, for example, 20 bytes (= 160 bits) as shown in FIG. At a predetermined position of each transmission unit block BL, a threshold value TH representing a quantization step width, DC component coefficient data DC, and AC component coefficient data AC are arranged, respectively.

As shown in FIG. 3, 6 bits are assigned to the threshold value TH and 10 bits are assigned to the DC component coefficient data DC. Also, 18 bytes (= 144 bits) are allocated to the coefficient data AC of the AC component. In this embodiment, a DCT block of (8 × 8) is cosine-transformed as a block transform code to generate coefficient data including a DC component DC and 63 AC components AC1 to AC63. Is quantized by a quantization step width corresponding to the threshold value TH, and the quantized output is subjected to variable-length encoding. In addition, an error flag indicating the presence or absence of an error is output one bit at a time for each byte (= 8 bits) of data. The error flag `0 'means no error and` 1' means error.

If there is no error in the data to be extracted, the Huffman code data is set to 1 as in the general method.
Bits are cut out, and a predetermined number of coefficient data of AC components are cut out and output. FIG. 4 shows an example of an output signal generated from the ECC circuit 5. Now, assuming that the coefficient data of the AC component of the block of interest is recorded from the leftmost bit of the third byte to the right, the frame decomposition circuit 6 starts to cut out data in order from the bit to the right. Since there is no error in the extraction target data, it is possible to extract the AC component coefficient data correctly. In this example, the Huffman table shown in FIG. 5 is used.

Next, the operation when there is an error in the cut-out data will be described. First, a conventional general method will be described. In the example of FIG. 6, an error flag is set in the data of the fourth byte. In the conventional method, when the extraction of the third byte of data is completed, the extraction of the AC component coefficient data of the block is terminated, and only the AC component coefficient data that has been picked up is output. In this example, the extraction of the coefficient data AC1 has not been completed at the time when the extraction of the third byte data has been completed. Therefore, all the coefficient data of the AC component cannot be picked up.

Next, the present invention will be described. FIG. 7 shows an example of the frame decomposition circuit 6 to which the present invention is applied. If an error occurs during cutout data, EC
The output signal of the C circuit 5 is supplied to the data extraction circuit 16. The data cutout circuit 16 changes the data of the fourth byte where the error has occurred from 0 to 255, and
It cuts out the coefficient data of the AC component as follows. The extracted 256 types of AC component coefficient data are stored in the AC component coefficient data memory 17. FIG. 9 shows an example of the contents stored in the coefficient data memory 17.

In the example shown in FIG. 8, data is cut out with the fourth byte of data set to 0.
In this example, as can be seen from a comparison with FIG. 4, since the data extraction has failed at AC2 and thereafter, not only AC1 but also data at AC2 and thereafter are random.

The output data of the AC component coefficient data memory 17 (the AC component coefficient data cut out in 256 ways) is supplied to the IDCT circuit 18. The IDCT circuit 18 converts the coefficient data of each AC component into an IDCT
Then, the image data is converted into 256 types of image data (each image data is an 8 × 8 block). The output data (256 kinds of image data) of the IDCT circuit 18 is
9. The correlation measuring circuit 19 calculates the correlation coefficient R by the following general equation for calculating a correlation.

R = Cov / (σX × σY) where R is two sets of data, X [1. . N], Y
[1. . N], σX is the standard deviation of X, σY
Is the standard deviation of Y, and Cov is the covariance of X and Y.

Actually, the vertical correlation coefficient and the horizontal correlation coefficient are calculated. Here, the vertical correlation coefficient is a correlation coefficient between pixels one pixel adjacent in the vertical direction on the image, and the horizontal correlation coefficient is a coefficient between pixels one pixel adjacent in the horizontal direction on the image. It is a correlation coefficient. The correlation measuring circuit 19 outputs the sum of the vertical correlation coefficient and the horizontal correlation coefficient as a correlation coefficient.

The output signal (256 types of correlation coefficients) of the correlation measuring circuit 19 is supplied to the maximum correlation determining circuit 20.
The maximum correlation determination circuit 20 selects the maximum correlation coefficient from the 256 types of correlation coefficients, and estimates that the combination of the AC component coefficient data that resulted in the correlation coefficient is the most correct. This selected AC component coefficient data memory 1
7 and output as an output of the read frame decomposition circuit 6.

Each data separated by the frame decomposing circuit 6 is supplied to a block decoding circuit 7, and original data and corresponding restored data are decoded for each block. That is,
The decoding of the variable length code, the inverse quantization, and the IDCT are performed.
Then, as shown in FIG. 1, the decoded data of the image from the block decoding circuit 7 is supplied to the distribution circuit 8. In the distribution circuit 8, the decoded data is separated into a luminance signal, a color difference signal U, and a color difference signal V. The luminance signal, the color difference signal U, and the color difference signal V are supplied to the block decomposition circuits 9, 10, and 11, respectively. Block decomposition circuits 9, 10 and 11
Converts decoded data in block order in raster scan order.

The output signals from the block decomposition circuits 9, 10 and 11 are supplied to error correction circuits 12, 13 and 14, respectively. The error correction circuits 12, 13 and 14 correct data having an error with peripheral data. Digital luminance signal Y from error correction circuit 12
Is taken out to the output terminal 15Y. Error correction circuit 1
The digital color difference signal U from 3 is taken out to an output terminal 15U. The digital color difference signal V from the error correction circuit 14 is taken out to an output terminal 15V.

FIG. 10 is a flowchart for explaining the processing of the above-described embodiment. First, from the result of the error correction processing (step 5a) in the ECC circuit 5, the presence or absence of an error in the data of interest is determined. If there is no error, data is cut out (one way) (step 22).

If there is an error, the target data is changed from 0 to 256, and 256 kinds of AC component coefficient data are cut out (step 23). The cut-out coefficient data of the 256 AC components is stored in the coefficient data memory 17 (step 24). Next, 25
Six types of data are decoded (step 25), and correlation coefficients are calculated for the decoded 256 types of data (step 26a). Then, the maximum value of the correlation coefficient is calculated (step 27).

It is determined in step 28 whether the calculated maximum value of the correlation coefficient is equal to or larger than a threshold value. If this is greater than or equal to the threshold value, coefficient data for generating such a correlation coefficient is retrieved from the coefficient data memory 17 (step 29). Otherwise, it is determined that correct data has not been obtained, and an error flag is set (step 30). Then, the coefficient data is subjected to block decoding processing by the block decoding circuit 7 (step 7a).

In summarizing the present invention shown in the above-described embodiment, the present invention uses a correlation coefficient of image data from among a plurality of sets of cut-out data including correct data. To select combinations with high On the other hand, there is a general method, in which an error occurs in a block, in which interpolation is performed using surrounding and past data, and a few methods that have already been proposed and which can be picked up without errors are proposed. The method of estimating the coefficient data of the lost AC component from the coefficient data of the AC component and the reconstructed data of the surrounding blocks can output data that does not break down with the surrounding blocks. However, in most cases, the value is far from the true restored data, and it is not possible to obtain a very good restored image for a block in which an error has occurred.

On the other hand, according to the method of the present invention, since data is always selected from a combination of a plurality of sets of cut-out data including correct data, there is a great possibility that the true restored data itself can be restored. Even if you do
There is a high possibility that data very close to the true restored data can be output. Further, according to the present invention, the amount of recorded information does not increase at all.

In the above-described embodiment, the byte on which the error flag is set is changed to all possible values of 0 to 255. May be regarded as having a 1-bit error, and the range of change may be limited. That is, IDCT and correlation coefficient are calculated in the same manner as described above for nine patterns obtained by adding data that does not invert bits at all to eight patterns obtained by inverting each bit of a byte having an error. If the number of relations is equal to or greater than a predetermined threshold, it may be estimated that the byte that generated the image data is correct. If the maximum correlation coefficient is smaller than the threshold value, the correction is performed as a block error. The fourth byte in the example of FIG. 8 is the second one-bit error.

Next, another embodiment of the present invention will be described. In the above-described embodiment, the correlation coefficient of the restored data of the target block is used as a parameter, and restored data having a large correlation coefficient is selected as having a high possibility of being correct. However, although the probability is low, there is a case where the correlation is not so large even in the true restored data. For example, in a flat image with a low luminance laser, the noise level is large, so that the correlation coefficient between neighboring pixels is small. In such a case, an erroneous determination may occur. In particular, when a large level difference occurs between adjacent blocks as a result, visual deterioration is conspicuous.

In order to solve such a problem, in another embodiment, a correlation coefficient between adjacent pixels including restored data of a block adjacent to the target block is obtained, and the correlation coefficient is determined to be the maximum. Is to be selected. There are two possible methods for using the restored data of the peripheral blocks for the correlation determination. One of them is to locally decode data of a peripheral block one by one at the time of decoding a block of interest and use the data. Another method is to first decode all error-free blocks, store the data in memory, and when decoding an erroneous block (ie, decoding using correlation determination), The data is read from the memory and used.

The first method has an advantage that the memory capacity is small, but when decoding a block having one error, it is necessary to decode eight peripheral blocks. Therefore, in the following example, the second method, that is, decoding is performed for all blocks having no error first, and the data is stored in a memory. And uses it.

FIG. 11 shows the overall configuration of another embodiment, and portions corresponding to FIG. 1 are denoted by the same reference numerals. As shown in FIG. 11, the output signal of the error correction circuit 5 is processed by being divided into two systems. First, in the system of the frame decomposition circuit 6, the block decoding circuit 7, and the delay circuit 31, data of a block containing no error is decoded. The delay circuit 31 is a memory that stores the output data of the block decoding circuit 7 for one frame (or one field). On the other hand, the delay circuit 32 and the frame decomposition circuit 3
3. In the system of the block decoding circuit 34, data of a block containing an error is decoded. The delay circuit 32 is a memory that stores the output signal from the error correction circuit 5 until the processing of the processing system of the block that does not include the error is completed.

The frame decomposing circuit 33 has the same configuration as the frame decomposing circuit 6 of the above-described embodiment (see FIG. 7). The frame decomposing circuit 33 performs a correlation measurement using the output signal of the delay circuit 31, that is, the decoded data of the peripheral block of the target block including the error, and determines the most correct data from the data including the error. The block is cut out and output to the block decoding circuit 34.
The block decoding circuit 34 decodes the data.

From the delay circuit 31, a decoded signal of a block containing no error is output and supplied to the distribution circuit 8. On the other hand, a decoded signal of a block containing an error is output from the block decoding circuit 34 to the distribution circuit 8.
The processing of the output signal of the distribution circuit 8 is the same as in the above-described embodiment.

FIG. 12 shows an example of the frame decomposition circuit 33. Since the configuration is the same as that of the frame disassembly circuit 6 in FIG. 7, a detailed description thereof will be omitted. The delay circuit 31 stores decoded data of an error-free block of the currently decoded frame (or field). The data is supplied to the correlation measuring circuit 19 of the block decomposing circuit 33, and the correlation measurement is performed using the data of the peripheral blocks.

As shown in FIG. 13, the decoded data of the peripheral blocks of, for example, two rows (or two columns) adjacent to the block boundary among the data of the adjacent blocks of the data of the (8 × 8) target block are subjected to the correlation measurement. Used for

FIG. 14 is a flowchart showing a procedure of processing according to another embodiment of the present invention. The same processing as in FIG. 10 described above is performed, except that in step 26b, the correlation coefficient is calculated including the decoded data of the adjacent block.

In another embodiment of the present invention, the correlation coefficient is calculated within a range including the decoded pixel of the target block and the adjacent block, and the output data is determined based on the calculated correlation coefficient. Therefore, the possibility of erroneous determination can be reduced, and as a result of the erroneous determination, the level difference from the adjacent block is large, and the possibility that visual deterioration is conspicuous can be reduced.

In the above-described embodiment, DCT is used as the block transform code. However, the present invention is not limited to DCT, but is applicable to general encoding using entropy code. The present invention can be applied not only to a case where the present invention is applied to a digital VTR but also to a case where various transmission paths are used.

[0047]

As described above, according to the present invention, there is a high possibility that a true decoded value or a decoded value very close to the true decoded value is obtained for the data of the block in which an error has occurred. For example, when the present invention is applied to a digital VTR and an error occurs in reproduced data, a good reproduced image can be obtained by using the method of the present invention.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of a digital VTR that implements an embodiment of an encoding method according to the present invention.

FIG. 2 is a schematic diagram illustrating a configuration example of a sync block.

FIG. 3 is a schematic diagram illustrating a configuration example of a transmission unit block.

FIG. 4 is a schematic diagram for explaining the operation of the frame decomposition circuit according to the present invention when there is no error in the extraction target data;

FIG. 5 is an example of a Huffman code.

FIG. 6 is a schematic diagram for explaining the operation of a conventional frame decomposition circuit when there is an error in cutout data.

FIG. 7 is a block diagram of an embodiment of a frame decomposition circuit according to the present invention.

FIG. 8 is a schematic diagram for explaining the operation of the frame decomposition circuit of the present invention when there is an error in the data to be extracted.

FIG. 9 is a diagram showing an example of the contents stored in an AC component coefficient data memory, which is a part of the frame decomposition circuit of the present invention.

FIG. 10 is a flowchart illustrating a process according to an embodiment of the present invention.

FIG. 11 is a block diagram showing a configuration example of a digital VTR embodying another embodiment of the present invention.

FIG. 12 is a block diagram showing a configuration of a frame decomposition circuit according to another embodiment of the present invention.

FIG. 13 is a schematic diagram showing data used for calculating a correlation coefficient in another embodiment of the present invention.

FIG. 14 is a flowchart illustrating a process according to another embodiment of the present invention.

[Explanation of symbols]

 5 Error Correction Circuit 6 Frame Decomposition Circuit 16 Data Extraction Circuit 17 Coefficient Data Memory 19 Correlation Measurement Circuit 20 Maximum Correlation Determination Circuit

Continuation of the front page (56) References JP-A-4-115628 (JP, A) JP-A-62-230266 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04N 7 / 24-7/68 H04N 1/41-1/419 H04N 5/91-5/956

Claims (5)

(57) [Claims] An image processing apparatus for processing coded image data which has been subjected to variable length coding for each block, comprising : an error flag obtaining means for obtaining an error flag indicating presence or absence of an error for each predetermined unit of the coded image data. For the encoded image data of the predetermined unit indicated by the error flag, all possible data values are given, and the given encoded image data is decoded according to the variable length encoding condition. Decoding means for decoding, and restoration data for each of the data values decoded by the decoding means.
And data, the correlation between the restored data of the neighbor blocks is maximum
And selection means for selecting the data values, the correlation, the condensate located around the block boundary
An image processing apparatus characterized by being obtained between original data . 2. The image processing apparatus according to claim 1, wherein said selecting means includes a step of selecting the restored data decoded by said decoding means and said adjacent block.
Using the block's restored data,
Determined the correlation coefficient between the restored data of two pixels adjacent in the vertical direction, a correlation coefficient between the restored data of two pixels adjacent in the transverse direction across the boundaries of the blocks it was determined An image processing apparatus characterized in that a maximum value is detected from the sum of correlation coefficients. 3. The image processing apparatus according to claim 1, wherein a combination of the plurality of data is limited to a 1-bit inverted error pattern. 4. The image processing apparatus according to claim 3, wherein the correlation coefficient obtained for the error pattern of one bit inversion is compared with a threshold value, and if the correlation coefficient is smaller than the threshold value, An image processing apparatus for performing processing for modifying an entire block. 5. An image processing method for processing coded image data which has been subjected to variable-length coding for each block, comprising : obtaining an error flag indicating the presence or absence of an error for each predetermined unit of the coded image data; A decoding step of giving all possible data values for the encoded image data of the predetermined unit indicated by the flag, and decoding the given encoded image data of each data value according to the condition of the variable length encoding; Restoring for each data value decoded in the decoding step
A data, and a step of correlation of the restored data of the neighbor block selects the data value as a maximum, the correlation, the condensate located around the block boundary
An image processing method characterized by being obtained between original data .