JPH0879754A - Image encoding device and image decoding device - Google Patents

Abstract

(57) [Summary] [Object] To obtain an image coding device capable of reducing distortion between adjacent blocks while preventing deterioration of original image data in a device which performs block coding. [Structure] The image data sent line by line is converted into data in units of blocks in an image data conversion unit 102, and a predetermined block encoding is performed in a compression unit 108. On the other hand, the original image data is the additional information extraction unit 104.
In (1), whether or not the boundary portion of each block is an edge is inspected, and edge information is created. This edge information is encoded by the encoding unit 106. The data adding unit 110 adds the coded data to the block code and outputs it as compressed data to the outside.
On the basis of this edge information, by not filtering the boundary portion that is the edge portion at the time of decoding, it is possible to reduce the blocking distortion and prevent the edge information of the original image data from being damaged. .

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding device and an image decoding device. In particular, the present invention relates to a device that divides image data into predetermined blocks and performs encoding / decoding for each block.

[0002]

2. Description of the Related Art In recent years, it has become necessary to handle a large amount of image information not only in various industrial fields but also in general homes and various entertainment facilities. To directly transmit and store such large-capacity image information consumes a large amount of communication line capacity and requires a large-capacity storage device.

Therefore, in general, image information is often transmitted and stored after being compressed and encoded by a predetermined compression and encoding method. As such a compression encoding method, a JPEG method for a still image or an MP for a moving image is used.
Various types such as the EG method have been proposed.

For example, the compression coding by the JPEG system is a coding system based on DCT. For example, the image information to be coded is divided into blocks of 8 × 8 pixels and coded for each block. Is done. The compression rate of such a compression encoding method is controlled by setting a conversion table.

FIG. 6 shows an example of a configuration block diagram of a conventional typical image coding apparatus. As shown in FIG. 6, according to the conventional image coding apparatus 10, the original image data is first supplied to the image data conversion unit 12. In the image data conversion unit 12, the image data sent line by line is converted into block unit data. Here, for example, an 8 × 8 block is used as the block.

The converted compression unit 18 performs block coding on the image information. This block coding is performed by, for example, DCT (Discrete Cosine).
Block coding using Transform etc. is applied. The data thus compressed is output from the image coding apparatus 10 as final compressed data. Next, the image decoding device 20 for decoding the compressed data created by the image coding device shown in FIG. 6 will be described. FIG. 7 shows an image decoding device 20 according to this embodiment.
The configuration block diagram of is shown. As shown in FIG. 7, the decompression unit 22 first decodes the compressed data into the original data. That is, block decoding is performed on the original block code to obtain a decoded image. A predetermined low-pass filter is applied to the decoded image in the smoothing unit 24 as necessary to remove various distortions and the like due to encoding. Then, the final reproduced image is output from the smoothing unit 24.

[0007]

However, according to this conventional block coding system, the coding is performed independently for each block without regard to the correlation between adjacent blocks. In other words, the continuity between adjacent blocks is lost. Therefore, JPE
Regardless of when the conversion table recommended as the G method is used, when the conversion table is set to have such a compression ratio, when the block-encoded image information is decoded, the edge portion of the block is Distortion often occurs. That is, the block-like distortion appears in the composited image information. Such a phenomenon remarkably appears particularly when the conversion table is set so that the compression rate becomes high.

In order to remove such block-like distortion, spatial filtering by a low-pass filter is generally performed as described above to make the distortion inconspicuous. If the so-called smoothing by such filtering is performed, the distortion of the edge can be removed, but there is a problem that the edge originally existing in the original image information is also smoothed and lost. For example, such an edge may be the outline of a character.
As described above, the conventional block coding method has a problem that distortion is likely to occur at an edge portion between adjacent blocks.

As a technique for reducing block distortion,
For example, in Japanese Patent Laid-Open No. 62-196990, when the block-coded image data is combined, a plurality of image data regarding each pixel of a pixel row extending over two blocks is used to From the image data in
It is disclosed that the image data relating to the pixels at the block boundary portion of the pixel column in the other block are extrapolated and predicted with respect to each other.

In Japanese Patent Laid-Open No. 5-176176, only the DC component is first extracted before the block-encoded image data is combined, and the image data is combined. A configuration is disclosed in which a correction value for reducing discontinuity of components is calculated based on the DC component.

According to these techniques for reducing the block distortion, the block distortion can be reduced, but the edge component originally included in the image data itself is also reduced, so that the image quality is significantly deteriorated. There's a problem.

The present invention has been made in view of the above problems, and an object thereof is to provide an apparatus for performing block coding,
An object of the present invention is to obtain an image encoding device capable of reducing distortion between adjacent blocks while preventing deterioration of original image data.

[0013]

The first aspect of the present invention is, in order to solve the above-mentioned problems, an image coding apparatus provided with block coding means for dividing image information into predetermined blocks and coding the blocks. The edge information output means outputs the edge information output means for extracting a boundary portion with an adjacent block and outputting information as to whether the boundary portion is an edge or not, and the block code output by the block encoding means. And an information adding unit for adding and outputting edge information to be output, the image coding apparatus.

A second aspect of the present invention is an image decoding apparatus for decoding the code output by the image encoding apparatus of the first aspect of the present invention, wherein the edge information and the block code included in the input code are combined. And a smoothing unit that performs smoothing by filtering only a block boundary portion that is not an edge based on the edge information.

A third aspect of the present invention is the image encoding device according to the first aspect of the present invention, which includes edge information encoding means for compressing and encoding the edge information output by the edge information output means, and the information adding means. Is an image coding apparatus characterized in that the compression-coded edge information is added to a block code outputted by the block coding means.

[0016]

The edge information output means in the first invention is
It is checked whether the boundary portion between blocks in block coding is an edge of an image. Then, since the information is added to the block code, it is possible to know in advance that the edge portion of the image and the boundary between the blocks match at the time of decoding according to this block code.

The smoothing means in the second aspect of the present invention does not perform filtering on the boundary portion of the image when the boundary portion is the edge portion of the image. Therefore, the edge portion of the image does not disappear.

The edge information coding means in the third aspect of the present invention compresses and codes the edge information to reduce the data amount thereof. Therefore, it is possible to prevent the compression rate of the entire code from decreasing.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT A preferred embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the structure of an image coding apparatus 100 which is a preferred embodiment of the present invention.
As shown in FIG. 1, the original image data is first supplied to the image data conversion unit 102 and the additional information extraction unit 104. In the image data conversion unit 102, the image data sent line by line is converted into block unit data. Here, as the block, for example, an 8 × 8 block is used as described above. On the other hand, in the additional information extraction unit 104, it is inspected whether the boundary portion of each block is an edge, and the additional information is created based on the inspection result.

The characteristic configuration of this embodiment is the additional information extraction unit 104. This additional information extraction unit 104
The operation will be described in detail with reference to FIG. FIG. 2 shows a schematic plan view centering on a predetermined block of interest C when the original image data is divided into blocks of 8 × 8 pixels. As shown in FIG. 2, the pixels of the target block C are Cij (i = 0 to 7, j = 0 to 0).
It is represented by 7). Further, the upper adjacent block of the target block C is called a block B, and the left adjacent block is called a block A. In the additional information extraction unit 104 according to the present embodiment, whether the boundary portion with the upper adjacent block B and the boundary portion with the left adjacent block A are edge portions as the additional information for the target block C, respectively. It is extracted whether or not. Note that the information about the lower adjacent block and the right adjacent block is (for the target block C)
Do not extract. For example, the boundary between the block and the lower adjacent block is extracted as additional information for the block below the target block C.

The information Hn (n = 0 to 7) about the boundary portion between the left adjacent block A is calculated by the following formula.

[0023]

If (| Cn0−An7) ≦ Th Hn = 0 else Hn = 1 (n = 0 to 7) Here, the threshold value Th is a threshold value for determining whether or not it is an edge. And an appropriate value is set according to the characteristics of the image. That is, the case where this Hn is 1 means that the density difference (luminance difference) of the image is large at the boundary with the adjacent block, and it means that there is a high possibility that an edge exists at the boundary part. . On the other hand, Hn
Is 0 means that the density difference (luminance difference) of the image was small at the boundary with the adjacent block,
It is judged that there is no edge at the boundary. Further, the information Vm (m = 0 to 7) about the boundary portion between the upper adjacent block B and the above-described Hn is also calculated by the following equation.

[0024]

## EQU00002 ## if (| C0m-B7m) .ltoreq.Th Vm = 0 else Vm = 1 (m = 0 to 7) As described above, the additional information for the target block C is extracted. When there is no upper adjacent block or left adjacent block, the additional information Hn and Vm are set to 1. The information extracted in this way is represented as follows, for example.

[Hn, Vm] = 00000000 000011111 Hn Vm In this way, the information extracted by the additional information extracting unit 104 is encoded by the encoding unit 106.
The information extracted by the additional information extraction unit 104 is 0.
Alternatively, it is expected that the sequence of 1 data rarely appears at random. Further, generally, it is considered that 0 is more frequently consecutive than 1 in many cases. That is, it is generally considered that there is little probability that the edge of the image information coincides with the block boundary of the block code.

Therefore, it is expected that the information extracted by the additional information extraction unit 104 can be greatly reduced in the data amount by compression encoding without impairing the information amount of the original data. The encoding unit 106 has a configuration provided from such a viewpoint, and applies encoding such as Huffman encoding to the extracted information to reduce the amount of data.

For example, [Hn, Vm] = 0000000
When data is represented as 000011110000001111110 ..., first, 0 data is 1
Since one data is continuous and one data is continuous four times, if each continuous length is taken, it is 11, 4, 6, 6, ...
・ ・ It becomes. According to the description method of JPEG, these are, for example, NNNN = 11, SSSS = 4, NNNN = 6, SS
It is represented as SS = 6. And these NNNN
Data is compressed by Huffman coding for / SSSS. In addition, this NNNN / SSS
If a table adapted to the occurrence frequency of S is used, high compression rate encoding can be realized.

In parallel with the encoding of the extracted information in the encoding unit 106, the block encoding of the image information is performed in the compression unit 108. This block coding is performed exactly like the conventional block coding.

As described above, in the present embodiment, the image information is compression-encoded in the compression unit 108, and the extraction information is compression-encoded in the encoding unit 106 using separate and independent conversion tables. This is to realize a high compression rate according to each data, but it may be preferable to use a common conversion table for both data from the viewpoint of hardware reduction.

The encoded extraction information obtained by the encoding unit 106 is added to the block code obtained by the compression unit 108. This addition is performed in the data addition unit 110, as shown in FIG. Then, the data to which the extraction information is added is output to the outside as final compressed data.

This data addition is simply added to the block code as it is, but it is also possible to adapt it to the existing standard. For example, in the JPEG image coding method, a segment is provided whose application can be freely determined by an application.
For example, in FIG. 3, “ISO DIS 10918-1
Requirements And Guideli
The diagram shown in B-16 of Annex B of "nes" is described. This figure is a diagram showing the structure of a marker segment in a comment segment. The markers and parameters in this figure are defined as follows.

COM: A comment marker, which is a marker indicating the start of a comment.

Lc: Indicates the comment segment length and the length of the comment segment.

Cm: Comment byte, the meaning of which depends on the application.

Therefore, by storing the extracted information in this comment byte, it is possible to implement the present invention while using the existing image format.

There are other locations where the extracted information can be stored. For example, FIG. 4 describes the diagram shown in B-17 of the above document. This figure is a diagram showing the structure of a marker segment in the application data segment. The markers and parameters in this figure are defined as follows.

APPr: Application data marker, a marker indicating the start of an application data segment.

Lp: represents the length of the application data segment and the length of the application data segment.

Ap: Application data byte, an arbitrary 8-bit value.

Therefore, it is also preferable to store the extraction information in this application data byte.

The feature of this embodiment is that information indicating whether or not the boundary between blocks is an edge is extracted and this information is added to the block code to form the final compressed data. That is. According to this compressed data, it is determined whether or not the boundary portion of the block is an edge. Therefore, it is possible to minimize edge distortion by performing more appropriate filtering during decoding.

Next, the image decoding device 112 for decoding the compressed data created by the image coding device shown in FIG. 1 will be described. FIG. 5 shows a configuration block diagram of the image decoding apparatus 112 according to the present embodiment.
As shown in FIG. 5, the decompressing unit 114 first decodes the compressed data into the original data. That is, block decoding is performed on the original block code in the same manner as in the conventional method, and a decoded image is obtained. Also,
The original additional information is obtained by subjecting the compressed additional information to Huffman decoding. That is, the original [Hn, V
m] data is obtained.

A feature of this embodiment is that when smoothing a decoded image that has been decoded, whether or not the smoothing is performed is determined based on the contents of the additional information. . That is, the smoothing unit 116 in FIG. 5 applies the low-pass filter to the decoded image. Whether or not to apply the low-pass filter to the pixels at the boundary portion is Hn or Vm of the additional information. Is performed only for pixels with 0. That is, Hn or Vm
It is determined that a boundary pixel having a value of 1 constitutes an edge, and no filtering is performed on this pixel. As a result, according to the present embodiment, it is possible to reduce block distortion based on block coding without erasing the edge portion of the original image.

It is preferable that the filtering is performed separately in the vertical direction and the horizontal direction. For example, attention block C
H2 of 0 is 0, the pixel C of the target block C is
Horizontal filtering is applied to 20. Similarly, when V2 of the target block C is 0, vertical filtering is applied to the pixel C02 of the target block C. As described above, by performing the filtering processing in the vertical direction and the filtering processing in the horizontal direction independently, and performing only the filtering corresponding to the direction of the edge corresponding to each boundary portion, the block distortion can be more efficiently reduced.

[0044]

As described above, according to the present invention, when the block boundary portion is a non-edge portion, the block boundary portion is smoothed, and when the block boundary portion is an edge portion. Since the block boundary portion is not smoothed, the image coding apparatus capable of reducing the block distortion can be obtained without unnecessarily deteriorating the image quality.

Since the edge information used for the smoothing process is also compressed, it is possible to prevent the compression rate from being lowered due to the addition of information.

[Brief description of drawings]

FIG. 1 is a configuration block diagram showing a configuration of an image encoding apparatus which is a preferred embodiment of the present invention.

FIG. 2 is an explanatory diagram illustrating an operation of the image encoding device according to the present embodiment.

FIG. 3 is an explanatory diagram of comment segments in a JPEG format.

FIG. 4 is an explanatory diagram of application data segments in a JPEG format.

FIG. 5 is a configuration block diagram showing a configuration of an image decoding apparatus according to the present embodiment.

FIG. 6 is a configuration block diagram of a conventional image encoding device.

FIG. 7 is a configuration block diagram of a conventional image decoding device.

[Explanation of symbols]

 100 image coding device 102 image data conversion unit 104 additional information extraction unit 106 coding unit 108 compression unit 110 data addition unit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI technical display location H04N 1/41 B

Claims (3)

[Claims] 1. An image coding apparatus provided with a block coding means for coding image information by dividing it into predetermined blocks and coding the same, by extracting a boundary portion between adjacent blocks and determining whether or not the boundary portion is an edge. Edge information output means for outputting the information, and information addition means for adding the edge information output by the edge information output means to the block code output by the block encoding means and outputting the added information. The image coding device. 2. The image decoding device for decoding the code output by the image coding device according to claim 1, further comprising: a decoding means for decoding the edge information and the block code included in the input code, An image decoding device comprising: a smoothing unit that performs smoothing by filtering only a block boundary portion that is not an edge based on the edge information. 3. The image coding apparatus according to claim 1, further comprising edge information coding means for compressing and coding the edge information output by the edge information output means, wherein the information adding means includes the compression coding. An image coding apparatus characterized in that the edge information thus generated is added to a block code outputted by the block coding means.