JP3125565B2 - Image coding 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 coding method for providing a special image effect.

[0002]

2. Description of the Related Art In recent years, the image quality of a picture coding method has been improved and the practical range has been reached. However, image quality is degraded when encoding an image to which a special effect is applied or a cinema-size image such as a movie, so that high image quality is required.

[0003] In the field of digital signal processing such as video discs and the field of digital broadcasting, it has been attempted to reduce the size of video software and devices by encoding video signals with high efficiency.
Experts Group) / ISO 11172 standard.

[0004] A conventional image coding method will be described below. FIG. 4 shows a block diagram of this conventional image encoding method. In FIG. 4, reference numeral 41 denotes a format converter, 42 denotes a frame rearranger, 43 denotes a block divider, 44 denotes a motion compensator, 45 denotes a two-dimensional orthogonal transform (hereinafter abbreviated as DCT), 46 denotes a quantizer,
47 is a variable length coding means, 48 is a signal multiplexing means, 49 is an inverse quantization means, 50 is an inverse DCT means, 51 is a frame accumulation and prediction means, 52 is a buffer means, and 53 is a rate control means.

The operation of the image encoding method configured as described above will be described below with reference to an NTSC video signal as an example.

First, an input NTSC video signal is converted by a format converter 41 into a source input format (SIF) which is a start format of the MPEG1 standard.
The video signal is converted into a video signal of the
After the order of the frames is rearranged in the order of receiving the processing, the block is divided into blocks of 8 × 8 pixels by the block dividing unit 43 and sent to the motion compensating unit 44 to obtain the motion vector information for each block. Compensation processing is performed. The video signal that has undergone the motion compensation processing is applied to the DCT unit 4.
5, the DCT coefficient is subjected to a DCT process to reduce the spatial redundancy, and the processed DCT coefficient is subjected to a non-linear quantization process at a given quantization width in order to further reduce the amount of information by a quantization means 46; The variable length coding means 47 receives a variable length coding process. The variable length coded video signal is multiplexed by the signal multiplexing unit 48 with the motion vector information obtained by the motion compensating unit 44 and output. The multiplexed information is buffered by the buffer means 52 and transmitted at a constant rate. At this time, the rate control means 53 for monitoring the filling rate of the buffer means 52 is
Quantization means 4 so that buffer 2 does not overflow or become empty
6 is controlled.

On the other hand, the video signal output from the quantization means 46 is subjected to inverse quantization processing by the inverse quantization means 49, and is subjected to inverse DCT processing by the inverse DCT means 50 to be restored to a restored prediction video signal. The restored video signal is sent to the frame accumulation and prediction means 51 and used for motion compensation processing and quantization noise reduction processing. According to such a procedure, the encoded video signal is output as a constant-rate bit stream.

As another conventional example, FIG. 3 is a schematic diagram in which a cinema-sized image is allocated to the current TV system.

FIG. 3 shows that the vertical boundary of the effective display section does not coincide with the block boundary.

[0010]

However, in the above-described conventional configuration, when encoding a video signal that has been subjected to mosaic processing, if the boundary of the mosaic processing does not coincide with the boundary of a block, the frequency component of the high frequency band is Blocks containing the edges of the mosaic boundary with the DCT. The block including the edge consumes a lot of bits, and has a problem that the image quality of the entire screen is deteriorated.

Also, in the conventional example in which a cinema-sized image is allocated to the current TV system, many bits are consumed and the non-display section is shifted to the non-display section at the boundary between the effective display section and the non-display section. There has been a problem that so-called mosquito noise is generated and is visually noticeable.

An object of the present invention is to solve the above-mentioned conventional problems, and to provide a high-quality image encoding method even when an image to which a special effect is applied or a cinema-size image is allocated to the current TV system. Aim.

[0013]

In order to achieve this object, an image coding method according to the present invention divides image data into a plurality of regions, and performs two-dimensional orthogonal transform, quantization, and coding on the blocks. the image encoding method of performing, block
Before the division processing, the number of pixels on each side is 2 n
(N is a natural number) for each rectangular area
This is an image encoding method in which the boundary of a flattened rectangular area is matched with the boundary of block division.

[0014]

According to the present invention, according to the above-described method, the start position of the boundary of the mosaic pattern coincides with the boundary of the block division, and one side of the block is divided into 2 n (n is a natural number) pixels. Therefore, it is guaranteed that the number of mosaic pattern boundaries in a block is always minimized.

In the most effective example of n = 3, the mosaic pattern boundary coincides with the block division boundary and there is no mosaic pattern boundary in the block, so that the block subjected to DCT is only a flattened DC component. , There is no need to encode each frequency component, and encoding is performed with a small amount of information. The amount of information saved in the mosaiced blocks is allocated to other parts of the image, improving the overall image quality of the screen.

[0016]

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

FIG. 1 is a block diagram showing an image encoding method according to the first embodiment of the present invention. In FIG. 1, 1 is a format conversion means, 2 is a mosaic processing means, 3 is a frame rearrangement means, 4 is a block division means, 5 is a motion compensation means, 6 is a DCT means, 7 is a quantization means, and 8 is a variable length. Encoding means, 9 is signal multiplexing means, 10 is inverse quantization means, 11 is inverse DCT means, 12 is frame accumulation and prediction means, 13 is buffer means, and 14 is rate control means.

In FIG. 1, the components other than the mosaic processing means 2 are the same as those of the conventional example, and the description of the same components will be omitted. The mosaic processing means 2 performs a smoothing process on a rectangular area from the coordinate position of the address information given the SIF format video signal. The size of the rectangular area to be smoothed is given by a mosaic coefficient.

The image coding method according to the present embodiment configured as described above will be described below in the case where n = 3 in both the horizontal and vertical directions where the operation is most effective. The address information given to the mosaic processing means 2 is synchronized with the coordinate information of the 8 × 8 pixel block boundary to be divided by the block dividing means 4, and the mosaic coefficient is also designated as a rectangular area of 8 × 8 pixels. The same area is cut out by the block dividing means 4 from the rectangular area subjected to the mosaic processing, and is subjected to DCT processing by the DCT means 6 in order to reduce the spatial redundancy. Since it has been subjected to the smoothing process, it has only a DC component.

As described above, according to the present embodiment, the effective DCT coefficient output from the DCT means 6 is obtained by matching the block boundary divided by the block dividing means 4 with the mosaic pattern boundary processed by the mosaic processing means 2. Only the DC component can be used, and the bit amount required for encoding this portion can be reduced. The reduced bits can be used for other parts, and the deterioration of the image quality can be suppressed by reducing the quantization width given to the quantization means 7.

FIG. 2 shows a current T showing a second embodiment of the present invention.
It is the schematic diagram which allocated the image of the cinema size to V system.

In FIG. 2, the vertical boundary of the effective display section coincides with the block boundary, and is encoded as in the first embodiment.

As described above, according to the present embodiment, when a cinema-sized image is allocated to the current TV system, the vertical boundary of the effective display section is made to coincide with the block boundary, whereby the first embodiment is different from the first embodiment. Similarly, it is possible not only to suppress the deterioration of the image quality but also to realize a good image in which mosquito noise does not occur on the non-display section side at the boundary between the effective display section and the non-display section.

In the first embodiment, the mosaic pattern processed by the mosaic processing means 2 has been described by using 8 × 8 pixels. However, even when a mosaic process smaller than 8 × 8 pixels is performed, the number of bits is smaller than in the past. Clearly, it can be encoded.

[0025]

As described above, the present invention relates to an image encoding method for dividing image data into a plurality of regions and performing two-dimensional orthogonal transformation, quantization, and encoding on the blocks.
A mosaic process for flattening an arbitrary rectangular area of the image data with the number of pixels of which one side is 2 n (n is a natural number) is included, and a mosaic pattern boundary of the mosaic process matches a boundary of the block division.

Also, image data of a cinema size having a screen aspect ratio different from that of the current TV system is divided into a plurality of areas, and the blocks are subjected to two-dimensional orthogonal transform, quantization,
In an image encoding method for performing encoding, it is possible to suppress deterioration of an image and perform encoding with good image quality by matching a vertical boundary of an effective display section of cinema-sized image data with a boundary of block division. it can.

[Brief description of the drawings]

FIG. 1 is a block diagram of an image encoding method according to a first embodiment of the present invention.

FIG. 2 is a schematic diagram in which cinema-sized images are allocated to a current TV system according to a second embodiment of the present invention.

FIG. 3 is a schematic diagram in which cinema-sized images are allocated to a conventional TV system of a conventional example.

FIG. 4 is a block diagram of a conventional image encoding method.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Format conversion means 2 Mosaic processing means 3 Frame rearrangement means 4 Block division means 5 Motion compensation means 6 DCT means 7 Quantization means 8 Variable length coding means 9 Signal multiplexing means 10 Inverse quantization means 11 Inverse DCT means 12 Frame accumulation And prediction means 13 buffer means 14 rate control means

Continuing on the front page (72) Hideaki Ogawa 1006 Kadoma Kadoma, Kadoma City, Osaka Prefecture Inside Matsushita Electric Industrial Co., Ltd. 58) Field surveyed (Int.Cl. ⁷ , DB name) H04N ^7/ 24-7/68 H04N 1/41-1/419

Claims (1)

(57) [Claims] 1. An image encoding method for dividing image data into a plurality of regions and performing two-dimensional orthogonal transformation, quantization, and encoding on the blocks, wherein the image data is divided into one side before the block division processing. of
For each rectangular area where the number of pixels is 2 to the nth power (n is a natural number)
An image coding method , comprising: flattening and making a boundary of the flattened rectangular area coincide with a boundary of the block division.