JPH1098620A - Picture processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute efficient compression, encoding through the use of correlation between small blocks by dividing picture data into blocks, redividing the blocks which are divided in accordance with picture information into the small similar blocks and encoding the obtained small blocks as one picture element. SOLUTION: Block picture data obtained by dividing original picture data are taken into a block picture buffer 10 and an encoder 12 obtains a representative value. A statistic computing element 15 obtains statistic from block picture data and the representative value. A selector 16 outputs the representative value to a block code buffer 17 when block picture data and the representative value are considered to be the same based on statistic quantity and a re-divider 13 re-divides block picture data when they cannot be considered to be the same. It is repeated until they are considered to be the same. Thus, original picture data can be compressed/encoded with high efficiency without losing global information by preserving the outlines of respective graphics as they are.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital copying machine for reading an image of a document by a scanner, converting the image into digital image data, and temporarily storing the image data to record the image, and transmitting the image data. Facsimile, or an image processing device used for a personal computer or the like that performs communication and processing of image data, particularly an image suitable for compressing and encoding the informational redundancy of digital image data of a color still image The present invention relates to a processing device.

[0002]

2. Description of the Related Art Conventionally, various types of image processing apparatuses have been proposed as image processing apparatuses for compressing and encoding the informational redundancy of digital image data of this type. So-called block truncation coding (block truncation coding) in which data is divided into blocks including a predetermined number of pixels and coding is performed for each block.
coding (BTC) method. In general, in the BTC method in which coding is performed for each block, the correlation between blocks is weakened because the coding is completed with the information in one block, and when the size of the block is set large to increase the coding efficiency, However, the correlation between the blocks cannot be sufficiently reflected in the code, and it is difficult to maintain the definition.

[0003] To cope with this problem, Japanese Patent Application Laid-Open No. 57-127363 discloses a block truncation method that can maintain the definition while improving the coding efficiency by utilizing the correlation between blocks. Encoding (DBTC)
A scheme has already been proposed. In this DBTC coding method, when pixel values in a block are substantially uniform, the average pixel value of the block is used as the only pixel value information, and pixel information corresponding to the pixel indicating the average pixel value is given to all pixels. The information is run-length coded (RL), and when the pixel value of the block of interest and the pixel value of the immediately preceding block are the same, the information is coded with a flag indicating the same, thereby improving the coding efficiency. The definition is maintained while taking into account the pixel values of adjacent blocks.

In the block run-length (BRL) coding system disclosed in Japanese Patent Laid-Open No. 59-44175, coding efficiency is increased by performing run-length (RL) coding on a block having specific pixel information. Is increasing.

Further, in the encoding method disclosed in Japanese Patent Laid-Open Publication No. 4-270473, a mode in which a block is approximated by one pixel value, and a plurality of blocks approximated by one pixel value are subjected to block run length (BRL). 2.) A mode for encoding, a mode for approximating a block with two pixel values, and a mode for approximating a block with four or more pixel values are provided to improve the encoding efficiency and image quality.

The Journal of the Institute of Image Electronics Engineers of Japan, Vol. 23, No. 1, 1994
In the coding method disclosed in “Image coding method based on segmentation and application to resolution conversion” on pages 3 to 10, the block is divided when the error between the original image and the code exceeds a predetermined value. In addition to maintaining the image quality, the regions having high correlation are integrated and chain-encoded after the division to secure the encoding efficiency.

[0007]

However, scanning by the block run length (BRL) disclosed in the above-mentioned JP-A-57-127363, JP-A-59-44175, and JP-A-4-270473. In the method that uses the correlation between blocks remaining in the direction, in a picture or the like having a strong correlation in the direction orthogonal to the scanning direction of the block, the correlation between the blocks cannot be reflected in the compression encoding, and the Is discontinuous, and the definition cannot be maintained.

The above-mentioned Journal of the Institute of Image Electronics Engineers of Japan, vol. 23, no.
994) As disclosed in “Image coding method based on segmentation and application to resolution conversion” on pages 3 to 10, when the error between the original image and the code exceeds a predetermined value, it is adapted by dividing the block. In the case of a method in which the block size is changed, the image quality is improved, but when compressing and encoding image data such as characters, ruled lines, or graphs composed of relatively simple elements,
Since it is necessary to have information in units of pixels such as bitmaps as codes, it is difficult to greatly reduce the amount of information per block except in a very small number of cases, such as when all are approximated by the same pixel value. There is a problem that the efficiency cannot be made very high.

Therefore, the above-mentioned Journal of the Institute of Image Electronics Engineers of Japan, Vol. 23, No. 1,
In the technique disclosed in (1994), it is conceivable to set a large block size in order to increase coding efficiency. However, when the block size is set to be large as described above, there is a problem in evaluating an error value for determining whether or not the error between the original image and the code exceeds a predetermined value. That is, when evaluating the error value, whether the error value is a value generated by noise dispersed in the entire block, or a value generated by image data such as a graphic that actually exists in a limited place, It is difficult to distinguish based only on the error value.

[0010] For this reason, there is a new problem that the reproducibility of an image such as a figure that actually exists locally is deteriorated in addition to being easily affected by noise dispersed in the entire block. In addition, when chain coding is performed by integrating block areas having high correlation, if the block size is set to a large value, the code amount increases accordingly, so that a significant improvement in coding efficiency cannot be expected. It also has

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems of the prior art, and an object thereof is to provide an image or the like having a strong correlation in a direction orthogonal to the block scanning direction. Image processing that can reflect the correlation between blocks in the compression encoding, is less susceptible to noise dispersed throughout the block, and maintains the definition while improving the encoding efficiency. It is to provide a device.

[0012]

According to a first aspect of the present invention, there is provided an image processing apparatus which divides input image data into blocks and encodes each block. A dividing unit that divides the input image data into blocks each including a predetermined number of pixels; a statistical information calculating unit that calculates statistical information of pixel data included in each of the blocks; and a statistical information obtained by the statistical information calculating unit. A determination unit that determines whether each of the blocks is a first block to be subdivided and encoded or a second block to be encoded without subdivision, based on Subdivision means for dividing one block into small blocks satisfying a predetermined condition; representative value calculating means for obtaining representative values for the small block and the second block; And having a coding means for coding for.

Further, in the configuration according to the second aspect, the subdivision means repeats division until each small block satisfies the predetermined condition, and the encoding means comprises: Any one of a plurality of encoding methods is applied based on the statistical information or the determination result of the determination unit.

Further, in the configuration according to claim 3,
3. The image processing device according to claim 2, wherein the subdivision unit subdivides the first block or the small block into one or a plurality of predetermined shapes, or subdivides into a predetermined number of small blocks. It is characterized by the following.

Further, in the configuration according to claim 4,
4. The image processing apparatus according to claim 3, wherein the adaptive encoding unit can change the number and type of selectable encoding units according to the size of the first block or the size of the small block. It is characterized by the following. (Operation) According to the image processing apparatus of the present invention, a block is re-divided into small blocks having a similar shape to a block in accordance with pixel information, and the obtained small block is coded as one pixel to perform coding. Efficient compression coding can be performed using the correlation between blocks.

In other words, the image processing apparatus according to the present invention changes the resolution in accordance with the pixel information in the block, and encodes the image at a resolution sufficient to reproduce the image and enhance the encoding efficiency, thereby achieving localization. This means that the correlation of the image in the region is reflected in the encoding.

Further, by setting the shape of the integrated block to a predetermined shape, the shape and size of the block can be clearly indicated only by adding a tag. Therefore, it is not necessary to describe the shape of the block with a chain code or the like. The amount can be reduced.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of an image processing apparatus according to the present invention. Reference numeral 1 denotes original image data of a document read by a scanner (not shown); Is divided into a predetermined size of n pixels (for example, m × n = 8 × 8 = 64 pixels), and 3 is sequentially divided into blocks 2 which are units of encoding in the original image data 1 This is a buffer for temporarily storing image data.

Reference numeral 4 denotes a block 2 of a predetermined size and a representative value calculator for calculating a representative value of a small block obtained by dividing the block 2 into smaller blocks as required.
An encoding unit including an encoder employing one or more encoding systems such as a BTC or DCT encoding system; and 5, a decoded image of the original image data 1 and a code obtained by the encoding unit 4. This is a statistic calculation means for calculating a predetermined statistic from.

Reference numeral 6 denotes a selection means for selecting whether or not to redivide the block 2 of a predetermined size based on the statistics obtained by the statistics calculation means 5, and 7 denotes the selection means 6.
A subdivision unit for subdividing the block 2 into a predetermined size when the subdivision is selected by the
Storage / transmission means for storing and transmitting image data encoded by

FIG. 2 is a more specific hardware block diagram of the image processing apparatus according to the present embodiment configured as described above.

In the figure, reference numeral 9 denotes a frame image buffer which is read by a scanner or the like (not shown) and stores one page of original image data 1 of an input document, and 10 denotes one block (for example, m × n = 8 × 8). = 64 pixels), a block image buffer for temporarily storing image data 2 for
A selector 1 determines whether the encoding target is the image data 2 of the size of one block as it is or a small block obtained by subdividing this one block into smaller blocks.
6 is a selection switch that switches based on the signal from the control unit 6.

Reference numeral 12 denotes an encoder for outputting data obtained by converting image data of a block to be encoded into one or more kinds of corresponding codes such as a BTC or DCT encoding method, and 13 denotes a subdivision of the block. Is an intermediate code buffer for temporarily storing one or more types of encoded image data of one block and image data obtained by decoding the encoded image data. .

Reference numeral 15 denotes a statistic calculator for comparing one or more kinds of coded image data obtained from the coder 12 with the original image data 2 to calculate a predetermined statistic; A selector for selecting appropriate image data from one or more types of encoded image data obtained from the encoder 12 based on the statistics obtained from the statistics calculator 15; Reference numeral 17 denotes a block code buffer for temporarily storing encoded data for one block, and reference numeral 18 denotes a frame code buffer for storing encoded image data for one page.

In the present embodiment, the shape of the small block divided by the block subdivider 13 is set so as to be only a similar shape to the shape of the block 2 extracted first.

Next, the operation of the image processing apparatus according to the embodiment of the present invention shown in FIG. 2 will be described with reference to the encoding flowchart of FIG.

First, in the image processing apparatus, one of the originals read and input by a scanner (not shown) or the like is read.
The original image data 1 for the page is taken into the frame image buffer 9 and stored. Next, since the original image data 1 stored in the frame image buffer 9 is encoded for each block, the original image data 1 for one page of the original document is encoded.
Is divided into, for example, 8 × 8 = 64 pixels of one block of pixel data 2 as shown in FIG. 4, and the divided one block of pixel data 2 is
And store it (step 1).

When the original image data 1 for one page of the original is divided into pixel data 2 for one block as described above, the total number of pixels in the vertical and horizontal directions of the original image data 1 for one page of the original is Is not divisible by the number of vertical and horizontal pixels corresponding to one block, that is, if the last number of pixels in the vertical and horizontal directions is less than a predetermined m or n (for example, m = n = 8), FIG. As shown in FIG. 5, a pixel having a predetermined pixel value is added to the last pixel in the vertical and horizontal directions, or the last pixel in the vertical or horizontal direction of the original image data 1 is copied by a predetermined number, By filling the pixel data 2 of the last block in the vertical direction and the horizontal direction, the original image data 1 for one page is completely divided into an integral number of blocks.

As described above, of the original image data 1 for one page stored in the frame image buffer 9,
For example, from the first block in the upper left,
The pixel data 2 for one block is stored in the block image buffer 1
0 and temporarily store it (step 1). Next, one block of pixel data 2 is stored in the block image buffer 10.
From the encoder 12 via the selection switch 11
A representative value for expressing the pixel data 2 for one block by one pixel value is obtained, and the pixel data 2 for one block is regarded as one pixel having the representative value as a pixel value (step 2). ).

Here, the one block of pixel data 2
As a representative value for representing, for example, the average value of all m × n pixel data of one block, or the average value of a plurality of pixel data located at a predetermined position in one block, or of one block An average value of one pixel data or a plurality of pixel data located at the center, or a median (median value) obtained by arranging m × n pieces of pixel data of one block in order of density is used.

To further explain, for example, when an average value of m × n pieces of pixel data of one block is used as a representative value for representing the pixel data 2 of one block, The average value p ₀ of all pixel data is
_Assuming that the density of each pixel in the block is a _ij , p
₀ = Σa _ij / (m × n).

Now, one block of pixel data 2 is shown in FIG.
If the average value p ₀ is used as a representative value for representing the pixel data 2, one block of pixel data 2 has a pixel density of the average value p ₀ . Will be considered a pixel.

Next, the encoder 12 encodes the block, in this case, as shown in FIG. 6, a representative value p ₀ representing one block of pixel data 2 by one pixel value.
A tag indicating the block size is generated as a code 20 (step 3), and the encoded pixel data 2 of one block is temporarily stored in the intermediate code buffer 14.

On the other hand, the statistic calculator 15 is obtained by the original image data 2 for one block temporarily stored in the block image buffer 10 and the encoder 12 temporarily stored in the intermediate code buffer 14. A statistic, for example, a mean square error, is obtained from the decoded image of the code 20 (step 4).

Here, assuming that the mean square error is E, the mean square error E is given by the following equation. E = Σ (a _ij −p ₀ ) ² where Σ is 1 to m for i and 1 to n for j.

Next, it is determined whether or not the block is subdivided by the selector 16 based on the statistics (step 5), and the original image data 2 for one block and the decoded image data are substantially the same. If it can be considered, the code 20 shown in FIG. 6 is sent to the block code buffer 17, and if it is determined that the block should be divided, the block is re-divided by the block re-divider 13 into a small block similar to the block, for example, The image is re-divided into と も に in both the vertical and horizontal directions (in the case of m × n = 8 × 8, 4 × 4 blocks) (step 6).

At this time, whether or not the block is subdivided by the selector 16 is determined, for example, by determining whether or not the mean square error E as a statistic is equal to or less than a predetermined value or exceeds a predetermined value. It is done by judgment. For example, as shown in FIG. 7 and FIG.
If there are a small number of pixels that can be regarded as noise in the image, the original image data 2 for one block
It is considered to be white or a uniform pixel of a given color.

Thereafter, as shown in FIG. 9, a representative value of the small block is obtained by the encoder 12, and the small block is regarded as one pixel having the representative value as a pixel value (step 7). In the same manner as above, the representative values p01 and p02 representing the pixel data of the re-divided one block by two pixel values, the tag indicating the block size, and the bitmap indicating the division state of the block are signed. 20 (step 3), and the encoded pixel data 2 of one block is temporarily stored in the intermediate code buffer 14.

Further, the original image data 2 of one block (4 × 4) after division temporarily stored in the block image buffer 10 and temporarily stored in the intermediate code buffer 14 by the statistic calculator 15. A statistic, for example, a mean square error is obtained from the decoded image of the code obtained by the encoder 12 (step 4), and the selector 16 selects whether or not to perform subdivision.

If it is determined that the block should be divided, as shown in FIG. 9, the block is re-divided by the block re-divider 13 into a small block similar in shape to the block, for example, a half (vertical and horizontal) size ( It is re-divided into 2 × 2 blocks (step 6).

When the block is subdivided into と も に in both the vertical and horizontal directions, the block composed of small blocks regarded as 2 × 2 pixels, more precisely, 2 × 2 pixels, is subjected to the next encoding. Then, the process returns to step 3, and the encoder 12 performs encoding by, for example, BTC to obtain a code 20 including two representative values and a bit map. In addition,
If the pixel size of the original image data is equal to the size of the small block, the number of selectable coding means is increased, and for example, DCT coding is included as a candidate, and the statistic obtained by the statistic calculator 15 is included. Is selected.

Steps 3 to 7 are repeated as described above, and when the original image data 2 for one block and the decoded image data can be regarded as substantially the same, or the pixel size of the original image data and the small block size become equal. Then, the subdivision of the block or the small block is stopped, and the encoding of one block is completed. Until all blocks of the original image data 1 for one page are completed,
The above processing is repeated (step 8).

(Modification) In the above embodiment, the case where the image processing apparatus is constituted by hardware including an encoder and a selector has been described. However, the present invention is not limited to this. Of course, it may be constituted by. Further, in the above embodiment, a configuration may be adopted in which a code obtained after a series of processes is subjected to lossless compression encoding.

[0044]

As described above, in the image processing apparatus according to the present invention, image data composed of relatively simple elements such as characters, ruled lines, and graphs, such as documents for business communication, and global information of the image can be obtained. Without loss, it is possible to perform compression encoding with extremely high efficiency while keeping the outline of each figure.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an embodiment of an image processing apparatus according to the present invention.

FIG. 2 is a hardware block diagram illustrating a more specific configuration of the image processing apparatus according to the embodiment.

FIG. 3 is a flowchart illustrating an operation of the image processing apparatus according to the embodiment.

FIG. 4 is an explanatory diagram showing one block.

FIG. 5 is an explanatory diagram showing a state in which one page of original image data is divided.

FIG. 6 is an explanatory diagram showing an encoding state of one block.

FIG. 7 is an explanatory diagram showing an example of an image of one block.

FIG. 8 is an explanatory diagram illustrating an example of an image of one block.

FIG. 9 is an explanatory diagram showing an encoding state of one block.

[Description of Code] 1 Original image data 2 Divided block 3 Buffer 4 Encoding means 5 Statistical calculation means 6 Selection means 7 Re-division means 8 Storage / transmission means

Claims (4)

[Claims] 1. An image processing apparatus that divides input image data into blocks and encodes the blocks, and divides the input image data into blocks including a predetermined number of pixels. A statistical information calculating means for obtaining statistical information of pixel data included in each block; and a first block to be subdivided and encoded based on the statistical information obtained by the statistical information calculating means. Determining means for determining whether or not there is a second block to be coded without re-dividing; re-dividing means for dividing the first block into small blocks satisfying a predetermined condition; Image processing comprising: representative value calculating means for obtaining a representative value for a small block and the second block; and coding means for coding the representative value. apparatus. 2. The re-division unit repeats division until each small block satisfies the predetermined condition. The encoding unit transmits the statistical information or the determination result by the determination unit. The image processing apparatus according to claim 1, wherein any one of a plurality of encoding methods is applied based on the information. 3. The method according to claim 2, wherein the subdivision unit divides the first block or the small block into one or more types of predetermined shapes, or divides the first block or the small blocks into a predetermined number of small blocks. The image processing device according to claim 2. 4. The encoding unit according to claim 1, wherein the number and types of selectable encoding systems can be changed according to the size of the first block or the size of the small block. Item 3. The image processing device according to Item 3.