JPS6356754B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an error concealment device in a coding system using orthogonal transformation, and relates to an error concealment device in a coding system using orthogonal transformation.
. . n) is replaced with a preset correction value αi to reduce image quality deterioration in the reproduced signal.

One of the orthogonal transforms used for encoding is the Hadamard transform, and in the present invention, a case where the Hadamard transform is used as the orthogonal transform will be described below. However, the present invention is not limited to Hadamard transform, but can also be implemented using orthogonal transform of values.

First, an example in which Hadamard transform is applied to an image signal will be described. Performing the Hadamard transform on the image signal means converting the time series data of the image signal into the frequency domain. n as the Hadamard transformation
Although there are the following (n: positive integer) transformations, in order to simplify the explanation, we will explain the fourth-order Hadamard transformation.

First, the image signal is sampled and divided into blocks of four sample points each. The values of these four points are X ₁ , X ₂ ,
If X ₃ and X ₄ and [H] is a 4th order Hadamard matrix, the 4th order Hadamard transformation can be expressed as the following equation [h ₁ , h ₂ , h ₃ , h ₄ ] is a fourth-order sequence transformed into the frequency domain by Hadamard transform.

Moreover, Hadamard matrix [H] can be expressed as follows.

When formula (1) is actually calculated using formula (2), each sequence hi (i=1, 2, 3, 4) becomes as follows.

h ₁ =X ₁ +X ₂ +X ₃ +X ₄ …(3) h ₂ =X ₁ +X ₂ −X ₃ −X ₄ …(4) h ₃ =X ₁ −X ₂ −X ₃ +X ₄ …(5) h ₄ =X ₁ −X ₂ +X ₃ −X ₄ (6) A case in which a color television signal is subjected to Hadamard transformation as an image signal will be explained. for example,
Four points d ₁ , d ₂ , d ₃ , and d ₄ shown in FIG. 1 are selected as sample points of the color television signal. Figure 1 shows an example of how to select sample points and block them in a color television signal.
-1 horizontal scanning line, A indicates the kth horizontal scanning line, C indicates the k+1st horizontal scanning line, and D indicates the k+2nd horizontal scanning line. Hadamard transformation is performed using sample points d ₁ , d ₂ , d ₃ , and d ₄ in FIG. 1 as one block. Assuming that the sample value of each sample point di is Xi, it is considered from equation (3) that h ₁ represents the DC component, that is, the average brightness of the block.
Furthermore, h ₂ is considered to represent a horizontal component from equation (4). Similarly, h ₃ is the diagonal component from equation (5),
From equation (6), h ₄ can be considered to represent a component in the vertical direction.

After each sequence hi, band compression can be performed by reducing the bit allocation to sequences with few components or sequences representing high frequency components to which visual sensitivity is low.

Next, a system configuration for performing band compression on an image signal and transmitting the image signal using Hadamard transform will be explained with reference to FIG.

FIG. 2 is a block diagram for explaining the configuration of a coding transmission system using Hadamard transform.

In FIG. 2, a color television signal arriving at an input terminal 1 is encoded into, for example, 8 bits by an A/D converter 2 and stored in a first memory 3. For example, as mentioned above, the sample values X ₁ of the four sample points shown in Figure 1,
Read out _{X 2 ,}
Find the fourth-order sequence hi (i=1, 2, 3, 4).

The first quantizer 5 quantizes each sequence to obtain Qihi. Qi (i=1, 2,...
4) is the quantization operator for each sequence. The first quantizer 5 allocates optimal bits to each sequence and reduces the number of bits of each sequence to be transmitted, thereby achieving band compression. For example, if the value of each sequence hi is represented by 8 bits, 32 bits are required to convert one block, but in the quantizer 5, 6 bits are used for Q ₁ h ₁ , 4 bits are used for Q 2 h 2, and 4 bits are used for Q ₂ h ₂ . ₃ h ₃ and
If 2 bits are assigned to each of Q ₄ h ₄ , it can be compressed to a total of 14 bits.

The transmission line encoder 6 encodes each quantized sequence.
Transmission line code 13 suitable for the transmission line that transmits Qihi
It modulates to. The above part shows the encoding side. The decoding side receives the transmission line code 13 transmitted from the encoding side, decodes it in the transmission line decoder 7, and reproduces the quantized signal Qihi of each sequence. Next, the second quantizer 8 obtains Q ^-1 _i Qihi. Q ^-1 _i is the inverse operator of Qi. The outputs hi'=Q ^-1 _i Qihi of the second quantizer 8 are each decoded into 8 bits. When band compression is performed, the decoded hi' has a different value from hi. The Hadamard inverse transformer 9 performs Hadamard inverse transform by the calculation shown in the following equation (7).

In equation (7), as can be seen from equation (2), [H ^-1 ]=[H] ...(8). Furthermore, the reproduced value Xi' is the reproduced value of the sample value Xi.

The reproduction value Xi' of the sample point obtained as the output signal of the Hadamard inverse transformer 9 is stored in the second memory 10. The second memory 10 reads out the reproduction value Xi' in synchronization with the synchronization signal of the color television signal,
The signal is converted into an analog signal by a D/A converter 11 and sent out from an output terminal 12 as a color television signal.

When transmitting image signals using a system having the above configuration, an error may occur in the transmission system. If an image signal is reproduced using a signal in which an error has occurred, the quality of the reproduced image will be significantly degraded.
Error correction encoding methods have been proposed to correct erroneous data. By adding redundancy to the original encoded signal, error correction codes
This is for error correction. Therefore, the number of bits required for transmission increases, and the effect of band compression using Hadamard transform is lost. Especially when the transmission system includes a magnetic recording system such as a VTR,
Relatively long burst errors such as dropouts also occur frequently. In order to correct a code that has caused such a burst error, a very large amount of redundancy must be provided, which further reduces the effectiveness of band compression.

On the other hand, error detection coding is a method that cannot perform error correction but can detect the occurrence of an error. Since the error detection code requires less redundancy than the error correction code described above, the effect of band compression is less likely to be impaired. As an error detection code, for example, there is a method of attaching a parity signal depending on whether a code word is an even number or an odd number.
In addition, in block coding, in which the code is divided into appropriate blocks and encoded, the number of “0” and “1” included in one block is determined, and during decoding, “0” and “1” are Errors can be detected by counting the number. In addition to these codes, error detection codes suitable for the transmission system used have been proposed.

When such an error detection code is used, it is important to consider how to reduce the influence of the error after the error is detected. Concealment refers to reducing the effect of erroneous data rather than correcting the erroneous data when an error occurs.

The present invention proposes a method for correcting errors in a coded transmission system using Hadamard transform, in which when an error occurs, the value of each sequence hi is replaced with a preset correction value αi.

FIG. 3 is a block diagram showing a decoding device according to the present invention.

A specific configuration of a decoding device for implementing the modification method according to the present invention in a transmission system using an n-order Hadamard transform (n=1, 2, 3, . . . ) will be described with reference to FIG. In FIG. 3, the decoder 7, second quantizer 8, Hadamard inverse transformer 9, second memory 10, and D/A converter 12 are the same as those shown by the same symbols in FIG. . In FIG. 3, a transmission path code 13 encoded according to the transmission path used is received at the input end 19, and a decoder 7 decodes the quantized signal Qihi (i=1, 2, ..., n) of each sequence. do. The quantized signal Qihi is sent to the second quantizer 8. The second quantizer 8 calculates the value of each sequence by performing an operation with the inverse operator Q ^-1 _i of the quantization operator Qi used on the transmitting side.
hi′=Q ^-1 _i Play Qihi. The quantized signal Qihi output from the decoder 7 is also sent to the error detector 20 for error detection. By using the above-mentioned error detection code as a transmission code, the error detector 20 detects the occurrence of an error. Further, when a magnetic recording system such as a VTR is used as a transmission system, a dropout detector can also be used as an error detector. When the error detector 20 detects an error, it generates an error signal ei according to the sequence hi in which the error occurs, and sends it to the switching circuit 21. The switching circuit 21 normally
The output signal 25 of the quantizer 8 is connected to the input signal 26 of the Hadamard inverse transformer 9, but when the error signal ei is received, the sequence indicated by the error signal ei is
The connection of the input signal 26 of the Hadamard inverse transformer 9 of hi is switched to the output signal 27 of the modification circuit 22. In the modification circuit 22, a modification value αi, which will be described later, is set in advance for each sequence. For example, if an error occurs in the sequence _h2 , the error detector 20 generates an error signal _e2 . Switching circuit 21
receives the error signal e ₂ and converts Hadamard inverse transformer 9
Among the input signals 26 of , the connection of the input signal corresponding to the sequence h ₂ ' is changed from the output signal 25 of the second quantizer to the output signal 27 of the modification circuit 22. By doing so, the Hadamard inverse transformer 9 reads the correction value α ₂ set in advance in the correction circuit 22 as the value of the sequence h ₂ ', instead of the signal in which the error has occurred. In other words, the Hadamard inverse transformer 9 receives the output hi' of the second quantizer 8 for sequences in which no errors occur, and receives the correction value αi set in advance in the correction circuit 22 for sequences in which errors occur. Then, perform the Hadamard inverse transformation operation described above,
Find the reproduction value Xi′. The reproduction value Xi' is stored in a memory 10, read out in synchronization with, for example, a synchronization signal of a color television signal, converted into an analog signal by a D/A converter 11, and sent out from an output end 12 as a color television signal.

When performing the above modification method, the modification value αi
The problem is what value to set.

The frequency of occurrence of each sequence value was measured when the 8th order Hadamard transform was performed on an actual color television signal. The measurements were performed on the color television signal at the eight sample points d _l1 , d _l2 , d _l+1 , shown in Figure 4.
₁ d _l+1,2 , d _l+2,1 , d _l+2,2 , d _l+3,1 , d _l+3,2 are used as one block to perform the 8th order Hadamard transformation. be. In Figure 4, F is the lth horizontal scanning line, K is the l+th horizontal scanning line.
1 and 2 indicate the (1+2)th horizontal scanning line, and (2) the (1+3rd) horizontal scanning line. Let the sample value at each sampling point be Xi, X ₁ = (d _l , 1), X ₂ = (d _{l +1} , 1), X ₃ = (d _{l +2} ,
1), X ₄ = (d _{l +3} , 1), X ₅ = (d _l , 2), X ₆ =
(d _l+1 , 2), X ₇ = (d _l+2 , 2), and X ₈ = (d _l+3 , 2), the eighth-order Hadamard transform can be expressed as follows.

Here, the Hadamard matrix [H] is as follows.

In the system shown in FIG. 2, the color television signal applied to the input terminal 1 is quantized to 11 bits by the A/D converter 2, and the Hadamard transformer 4 performs the 8th order Hadamard transform shown in equation (9), The output was measured for each sequence, and the frequency of occurrence of each value of the output level was determined. FIG. 5 shows the measurement results of the cumulative distribution of output levels for the sequence h ₁ and FIG. 6 for the sequences h ₂ to h ₈ . Note that the distribution of the frequency of occurrence at each level can be determined by differentiating the cumulative distribution curve.

From Figure 5, it can be seen that the output level of sequence h ₁ is distributed almost uniformly across each value, and from Figure 6, the output levels of sequences h ₂ to _{h 8} are concentrated near 0. . This is because the reproduction screen of a color television signal has many uniform components, that is, low frequency components, and relatively few high frequency components included in fine patterns.

Correction value αi of sequence hi (i=1, 2,...8)
As such, it is appropriate to take the value where the frequency distribution of the output level of each sequence is concentrated.

As shown in FIG. 6, the frequency of occurrence of the sequence hk (k=2, 3, . . . 8) is concentrated near the output level 0, so 0 may be taken as the correction value αk. Further, when 0 is used as the modification value, the system configuration becomes simple. Regarding the sequence _h1 , as shown in Figure 5, the output level is distributed almost uniformly among the values, so it is appropriate to take the middle value of the possible range of the output level as the correction value _α1 . be. In the case shown in Figure 5, the values of the sequence _h1 are 0 to
2047, the intermediate value 1024 may be set as the correction value _α1 .

Although the above results are obtained when an 8-order Hadamard transform is performed, the same is generally true when an n-order Hadamard transform (n: a positive integer) is performed.

In a signal processing method using the Hadamard transform, in order to correct the value of a sequence in which an error has occurred using an error correction code, a large degree of redundancy must be provided, resulting in poor processing efficiency. Therefore, errors are detected using an error detection code that has less redundancy than an error correction code, and the value of the sequence in which the error occurred is replaced with a preset correction value. By applying the retouching method according to the present invention in which an image signal that undergoes orthogonal inverse transformation is reproduced using values of sequences that have not been used, image quality deterioration due to errors in reproduced images can be effectively reduced.
The efficiency of signal processing can also be increased.

In particular, by selecting the value with the highest probability of appearing in each sequence as a modification value, the probability that the modification value will match the correct value or its vicinity increases, and the effect of errors on the playback screen can be minimized. . In color television signals,
In a sequence of h ₂ or more, the value with the highest probability of appearance is close to 0, and by taking 0 as the correction value, the actual hardware configuration is also very simple.

[Brief explanation of the drawing]

Figure 1 is a diagram showing an example of how to select sample points in a color television signal and block it. Figure 2 is a block diagram for explaining the configuration of a coding transmission system using Hadamard transform. , FIG. 4 is a block diagram of a decoding device showing the first embodiment of the present invention, and FIG. 4 is a diagram showing an example of how to select sample points and block them when performing the eighth-order Hadamard transform on a color television signal. is a graph showing the cumulative distribution of the output level of sequence _h1 when the 8th Hadamard transform is applied to the color television signal, and Fig. 6 is the cumulative distribution of the output level of the sequence h1 when the 8th Hadamard transform is applied to the color television signal. _{2 is a graph showing the cumulative distribution of output levels from h8} to _h8 . 2...A/D converter, 3...Memory, 4...Hadamard transformer, 5...quantizer, 6...encoder, 7...decoder, 8...quantizer, 9...Hadamard inverse transformer, 1
0... Memory, 11... D/A converter, 20... Error detector, 21... Switching circuit, 22... Correction circuit.

Claims (1)

[Claims] 1. Divide the image signal into blocks, perform n-th orthogonal transformation on the blocks, transmit the sequence hi (i = 1, 2,...n) obtained as the orthogonal transformation output, and receive it. In a device that reproduces an image signal by performing orthogonal inverse transform on the received signal, the device includes an error detector that detects errors in each sequence of received blocks, and a correction circuit that presets correction values for each order of the sequence. , a switching circuit that replaces the value of the sequence in which the error occurred with a corrected value, and an inverse transformer that performs orthogonal inverse transformation using the corrected value of the sequence in which the error occurred and the value of the sequence in which no error occurred. An orthogonal transform decoding device characterized by: 2. Claim 1, characterized in that each sequence is provided with a modification circuit that sets a value with the highest probability of occurrence in that sequence as a modification value.
The orthogonal transform decoding device described in . 3. The orthogonal transform decoding apparatus according to claim 1, further comprising a modification circuit that sets 0 as modification values for sequences other than at least the lowest order sequence.