JP2643636B2 - Signal 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 a signal processing method for encoding and decoding a video signal by band compression involving block processing such as DCT.

[0002]

2. Description of the Related Art In recent years, transform coding such as DCT has become the mainstream in band compression. These use the information of the original image as 1 /
This is a method of highly efficient compression to about 10 to 1/40.

[0003]

However, as the compression ratio increases, so-called block distortion occurs, in which the steps between blocks are conspicuous. For example, in units of 8 pixels vertically and horizontally, DC
When a block of T is formed, a level difference is generated in a vertical and horizontal eight pixel cycle, resulting in an unsightly decoded image.

[0004]

SUMMARY OF THE INVENTION The present invention provides a method for reproducing a video signal .
Is bandwidth compression coding the number of pixels as one block
A video signal processing method that, when decoding ,
Change in the level of the pixel at the boundary between the blocks and the previous
Comparison with the change in the level of the pixel near the boundary ,
Result of the comparison, the case where the change in the level of the boundary pixels between adjacent blocks changes greater than the level of the pixels of the vicinity of the boundary within the block, so that the change in the level of the pixel of the boundary is reduced, Pixel level of the boundary,
Alternatively, the pixel level at the boundary and the pixel level near the boundary are corrected and then output.

[0005]

According to the present invention, when a step between blocks occurs, the step between blocks is smoothly corrected without deteriorating the original resolution of the image, and the block distortion can be significantly reduced visually. is there.

[0006]

An embodiment of the present invention will be described below with reference to the drawings.

[0007] Now, consider that one block consists of 8 x 8 pixels. FIG. 1 schematically shows the data in the horizontal direction. Here, each of the pixels 1 to 8, 9 to 16, and 17 to 24 is one block.

FIG. 1 shows an example in which a step occurs between pixels 8-9 and 16-17 at the block boundary. In this case, the pixel is output after being corrected to the pixel indicated by the black circle.

A method for obtaining a correction value will be described. Assuming that the level of the pixel 8 is a and the level of the pixel 9 is b, the pixel level of 8 is a- (ab) / K. The pixel level of 9 is b- (ba) / K. Here, K takes a value of about 3 to 5.

As a result, it is possible to reduce the level difference between the blocks and to obtain a great improvement in image quality visually.

However, when an image having a large data difference between blocks is actually encoded and decoded, it is not preferable to correct the image because the outline is blurred.

Therefore, if the difference | ab | within a block is equal to or larger than a predetermined amount, it is desirable that the original image is determined to be such an image and no correction process is performed.

That is, the above-described correction is performed only when | ab | is smaller than a predetermined amount. Here, the predetermined amount is
It is set to around 10 at the time of 8 bits.

FIG. 2 shows a second embodiment. This is to perform processing in consideration of both the level between two pixels and the level between pixels between blocks in the same block at the block boundary.

Assuming that the levels of the pixels 7, 8, 9, and 10 are a, b, c, and d, respectively, when block distortion is conspicuous, | ab | and | cd | This is when c | is large. At this time, as shown by the dotted line in FIG. 2, pixel 7 is a- (bc) / m pixel 8 is b- (bc) / n pixel 9 is c- (cb) / n By correcting the pixel 10 to d− (c−b) / m, the step between blocks can be made smooth. Here, n takes a value of about 4, and m takes a value of about 8.

In this case, only the pixels 8 and 9 may be corrected as described in the first embodiment. In addition, the pixel differences | ab | and | cd |
If it is larger than (approximately 8 in bit quantization), there is no need to perform correction because the block step is originally inconspicuous.

When the pixel difference between blocks is equal to or larger than a predetermined value (for example, about 10 for 8 bits), it is considered that there is a difference in the original image and no correction is performed in the same manner as in the first embodiment. good.

Next, the pixels 15, 16, 17, and 18 will be described. Let each level be e, f, g, h. When the difference | ef | or | gh | between two pixels in adjacent blocks is larger than the difference | fg | between blocks, it is considered that the step between blocks is caused by the original signal. Good, no need to correct.

In this case, | ef |> | fg |> | gh |, so there is no need for correction.

That is, the conditions for performing the correction in the second embodiment are as follows: The block pixel difference (| ab |,
| Cd |,) is a predetermined value (about 8 in the case of 8 bits)
Smaller than 2. The pixel difference between the blocks is larger than the pixel difference in either of the adjacent blocks (| bc |> | ab | and |
bc |> | cd |) 3. The pixel difference between blocks is a predetermined value (10 in case of 8 bits).
Degree) is to satisfy all or any of these three conditions.

Here, the above-described correction is performed in all or any of the horizontal, vertical, and time axis directions of the screen.

Further, the predetermined value is not limited to the value described in the embodiment. Furthermore, the predetermined value may be adaptively changed according to the screen, for example, according to the noise amount, the recording / reproducing mode, or the like.

In the embodiment, the correction of 2 to 4 pixels at the block boundary has been described, but the number of pixels is not limited to this.

The block size is not limited to 8 × 8, but may be any size such as 4 × 4.

Further, the correction itself according to the present invention is turned ON,
It may be turned off and controlled based on information such as a compression ratio and motion detection.

[0026]

[0027]

The present invention as described above, according to the present invention is, movies were bandwidth compression coding a plurality of pixels of the image signal as one block
A method of processing an image signal, the method comprising the steps of:
Change in the level of the pixel at the boundary between and the
By comparing with the change in the level of the pixel near the boundary , the ratio
Result of compare, the case change in the level of said boundary pixels between adjacent blocks changes greater than the level of the pixels of the vicinity of the boundary within the block, so that the change in the level of the pixel of the boundary is reduced, The pixel level of the boundary, or
Is configured to output after correcting the pixel level at the boundary and the pixel level near the boundary.

Thus, when a step between blocks occurs, the step between blocks is smoothly corrected without deteriorating the original resolution of the image, and the block distortion can be significantly reduced visually. Can be obtained.

[Brief description of the drawings]

FIG. 1 is a schematic pixel level diagram illustrating a first embodiment of the present invention.

FIG. 2 is a schematic pixel level diagram illustrating a second embodiment of the present invention.

Claims (6)

(57) [Claims] A plurality of adjacent pixels are defined as one block.
To process video signals that have been band-compressed and coded.
When decoding , the level of the pixel at the boundary between adjacent blocks changes.
And the level change of pixels near the boundary in the block.
Compares the reduction, the result of the comparison, the case where the change in the level of the boundary pixels between adjacent blocks changes greater than the level of the pixels of the vicinity of the boundary within the block, the levels of the pixels of the boundary Pixel level at the boundary, such that the change is small ,
Alternatively , the signal level is output after correcting the pixel level at the boundary and the pixel level near the boundary. 2. The level of a pixel near a boundary in a block.
The change is based on the pixel located at the boundary in the same block and the block.
This is a level change with a pixel not located on the boundary within the lock.
2. The signal processing method according to claim 1, wherein: 3. A method in which a plurality of adjacent pixels are combined into one block.
To process video signals that have been band-compressed and coded.
Te, upon decoding, Le between two pixels in the same block of the block boundary
The change in the bell is less than or equal to a predetermined amount, and
The change in level between two pixels between blocks
If it is more than the level change between two pixels in the lock,
The change in level between the two pixels between the blocks will be small
As described above, the pixel level at the block boundary or the block boundary
Pixel level of the field and pixel level near the block boundary
Output after correcting the signal
Law . 4. The signal processing method according to claim 1 , wherein compensation is performed such that the level difference is reduced only when a change in level between pixels between blocks is equal to or less than a predetermined amount. 5. A signal reproducing apparatus for reproducing a video signal.
5. The video signal according to claim 1, 2, 3, or 4,
Characterized in that it is processed and output by the signal processing method of
No. playback device . 6. A signal processing device for processing a video signal.
5. The video signal according to claim 1, 2, 3, or 4,
Shin, characterized in that processes and outputs in the signal processing method
No. processing unit .