JP3488672B2 - IMAGE PROCESSING APPARATUS, IMAGE PROCESSING METHOD, AND COMPUTER-READABLE RECORDING MEDIUM RECORDING PROGRAM FOR CAUSING COMPUTER TO EXECUTE THE METHOD - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention divides a multi-valued image into blocks of coding units and encodes image data in the divided blocks into fixed length data, an image processing method, and the same. The present invention relates to a computer-readable recording medium in which a program for causing a computer to execute the method is recorded, and in particular, to perform fixed-length encoding that reflects partial characteristics of an image in encoded data without causing a decrease in encoding efficiency. The present invention relates to an image processing device, an image processing method, and a recording medium capable of performing the above.

[0002]

2. Description of the Related Art In recent years, image processing apparatuses such as facsimiles, copiers and digital cameras have been improved in image quality year by year by increasing resolution and increasing the number of gradations. Since the information amount of the image is increased, the storage capacity for storing the image data is increased.

For example, a black and white image with two gradations (binary image)
When a black and white grayscale image having 256 gradations is used, the amount of information is 8 times, so that a memory for storing such image data is required 8 times, resulting in an increase in equipment cost.

For this reason, recently, an encoding technique for compressing and encoding image data has been in the limelight. Specifically,
As this encoding technique, a variable length encoding technique for converting image data into variable length encoded data is known.

For example, in Japanese Unexamined Patent Publication No. 2-122767, a plurality of transform coefficients are obtained by performing an orthogonal transform such as a two-dimensional discrete cosine transform on each block of an image, and the transform coefficients are coded to obtain a variable length code. An encoding / decoding method of an image signal configured to perform encoding is disclosed.

Since the conventional variable-length coding technique represented by this conventional technique has high coding efficiency and is reversible, it is possible to perform coding suitable for an image forming apparatus such as a facsimile.

[0007]

However, in such a variable length coding technique, since the position of the image before coding is unknown in the coded state, it is not possible to reproduce only the image of an arbitrary part. Therefore, this variable length coding technique
It cannot be applied to an image forming apparatus that involves image processing such as a copying machine.

Therefore, in the case of an image forming apparatus such as a copying machine, a fixed length coding technique is required in which the coding result of one block of an image is always fixed length, but an efficient fixed length coding technique. Is not known.

Specifically, in the conventional fixed length coding technique, it is general to take out the upper 2 bits from each pixel of, for example, 8 bits and compress it into a 1/4 image. According to this, the partial characteristics of the image cannot be reflected in the code data.

In other words, the conventional fixed length coding technique cannot distinguish visually important image areas from other image areas, resulting in deterioration of image quality. In addition, in order to improve the image quality, the number of bits assigned to one pixel is inevitably increased, which leads to a decrease in coding efficiency.

For these reasons, it is extremely important to efficiently realize the fixed-length coding technique capable of reflecting the partial feature of the image in the code data without lowering the coding efficiency. It has become a challenge.

The present invention has been made in order to solve the above-mentioned problems, and performs fixed-length coding in which partial characteristics of an image are reflected in code data without causing a reduction in coding efficiency. An object of the present invention is to provide an image processing apparatus, an image processing method, and a computer-readable recording medium in which a program that causes a computer to execute the method is recorded.

[0013]

[Means for Solving the Problems]
In order to achieve the object, an image processing apparatus according to the invention of claim 1 divides a multi-valued image into blocks of coding units, and encodes image data in the divided blocks into fixed length data. In which the block is subdivided into a plurality of small blocks, an extracting unit that extracts a feature amount from the image data in the small blocks redivided by the subdividing unit, and an extracting unit that extracts the characteristic amount. An extracting unit that determines the number of bits to be assigned to each small block based on the feature amount of the image data in each small block, and encodes the image data in the block into fixed-length data. Is a low-pass filter that calculates the average of each pixel of the image data in the small block to extract the pixel average, and A high-pass filter for obtaining a difference between a large pixel value and a minimum pixel value is applied to extract edge information, and the encoding means sets a conversion mode based on the presence or absence of edge information of each small block forming the block. A conversion mode determining unit for determining, a conversion table in which the number of bits allocated to each small block and the conversion mode are associated with each other, and a fixed unit that encodes image data in the block into fixed-length data based on the conversion table. And a long coding means.

According to the invention of claim 1, the block is subdivided into a plurality of small blocks, the feature amount is extracted from the image data in the subdivided small blocks, and the image data in each of the extracted small blocks is extracted. Based on the feature amount, we decided to encode the image data in the block into fixed length data,
It is possible to perform fixed-length coding that reflects the partial characteristics of an image in code data without causing a reduction in coding efficiency. Further, a conversion table in which the number of bits allocated to each small block and the conversion mode are associated with each other is provided, the conversion mode is determined based on the presence or absence of edge information of each small block forming the block, and the conversion table is determined based on the conversion table. Since the image data in the block is encoded into the fixed length data, the conversion table can be used to efficiently perform the fixed length encoding.

The image processing apparatus according to a second aspect of the present invention is the image processing apparatus according to the first aspect of the invention, wherein the conversion table includes the number of bits to be assigned to the small blocks that do not have the edge information and the small blocks that have the edge information. It is characterized in that the number of bits to be allocated is stored for each conversion mode.

According to the second aspect of the present invention, the number of bits assigned to the small block having no edge information and the number of bits assigned to the small block having edge information are stored in the conversion table for each conversion mode. It is possible to assign a different number of bits depending on the presence or absence of an edge, and thus perform fixed length coding in consideration of the edge.

The image processing apparatus according to a third aspect of the invention is the image processing apparatus according to the second aspect of the invention, in which the number of bits to be assigned to the small blocks not having the edge information in each conversion mode and the small blocks having the edge information are assigned. The sum with the number of bits is characterized by having a predetermined bit length.

According to the third aspect of the present invention, the sum of the number of bits assigned to the small block having no edge information and the number of bits assigned to the small block having edge information in each conversion mode has a predetermined bit length. Therefore, the code data can always have a fixed length.

Further, in the image processing apparatus according to the invention of claim 4, in the invention of claim 2 or 3, the conversion table is arranged such that when the small block has edge information, a pixel position of a pixel forming an edge. The number of bits allocated to pixels forming the edge and the number of bits allocated to pixels not forming the edge are stored.

According to the invention of claim 4, when the small block has edge information, the pixel position of the pixel forming the edge, the number of bits to be allocated to the pixel forming the edge, and the pixel not forming the edge are allocated. Since the number of bits is stored, it is possible to perform fixed-length coding that reflects the pixel positions of the pixels forming the edge in the small block.

Further, in the image processing apparatus according to the invention of claim 5, in the invention of claim 2, 3 or 4, the fixed length coding means is configured to: if the small block has edge information, the small block. The bit string forming the pixel value of each pixel is truncated to the number of bits allocated to the pixels forming the edge or the bit string allocated to the pixels not forming the edge.

According to the fifth aspect of the present invention, when the small block has edge information, the number of bits assigned to the pixels forming the edge or the bit string forming the pixel value of each pixel in the small block is not formed. Since the bit string assigned to the pixel is truncated, the data can be compressed quickly.

The image processing apparatus according to the invention of claim 6 is the image processing apparatus according to any one of claims 2 to 5, wherein the fixed-length coding means includes: when the small block has no edge information;
The image data in the small block is cut into a predetermined bit string based on the pixel average extracted by the extracting unit.

According to the invention of claim 6, when the small block does not have edge information, the image data in the small block is truncated to a predetermined bit string based on the average of the extracted pixels. It is possible to efficiently perform data compression of pixels that do not form edges.

An image processing apparatus according to a seventh aspect of the invention is an image processing apparatus for dividing a multi-valued image into blocks of coding units and coding the image data in the divided blocks into fixed length data. A multiplex division unit that recursively multiplexes the block into small blocks having a hierarchical structure, and an extraction unit that extracts edge information from the image data in the small blocks that are multiplexed by the multiplex division unit.
Encoding means for encoding the image data in the block into fixed length data based on the feature amount of the image data in each small block extracted by the extracting means, wherein the extracting means comprises the small block A high-pass for obtaining the difference between the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block while applying a low-pass filter for obtaining the average of each pixel of the image data in The edge information is extracted by applying a pass filter, the encoding means determines a conversion mode based on the presence or absence of edge information of each small block forming the block, and a conversion mode determining means for each small block. A conversion table that associates the number of allocated bits with the conversion mode, and a fixed table that encodes image data in the block into fixed-length data based on the conversion table. Characterized in that and a length coding means.

According to the seventh aspect of the present invention, the block is multiplexed into small blocks, the edge information is extracted from the image data in the small blocks, and the small blocks having the edge information are recursively re-divided. After repeating the process, the image data in the block is encoded into fixed-length data based on the feature amount of the image data in the small block. Further, it is possible to improve the image quality by reflecting it in the code data.

Further, a low-pass filter for obtaining the average of each pixel of the image data in the small block is applied to extract the pixel average, and the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block are extracted. Since we decided to apply a high-pass filter that calculates the difference of
Fixed-length coding that considers edges such as contour lines can be performed.

Further, a conversion table in which the number of bits allocated to each small block and the conversion mode are associated with each other is provided, and the conversion mode is determined based on the presence / absence of edge information of each small block forming the block. Since the image data in the block is encoded into the fixed length data based on the table, it is possible to efficiently perform the fixed length encoding using the conversion table.

Further, the image processing apparatus according to the invention of claim 8 divides the CMYK color system color image into blocks of encoding units, and encodes the image data in the divided blocks into fixed length data. In the image processing device, a feature is defined by subdivision means for subdividing the block into a plurality of small blocks for each component of C, M, Y, and K, and image data in the small blocks subdivided by the subdivision means. Based on the feature amount of the image data in each small block extracted by the extractor, and the C, M, Y, and K data in the block are encoded into fixed-length data. Encoding means, and the extracting means extracts a pixel average of each of C, M, Y, and K components by applying a low-pass filter that obtains an average of each pixel of the image data in the small block. In both cases, a high-pass filter that finds the difference between the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block is applied to extract edge information of each of C, M, Y, and K components, and the code The conversion unit determines a conversion mode based on the presence or absence of edge information of each small block forming the block, and a conversion table that associates the number of bits allocated to each small block with the conversion mode. And C, M, Y, K for forming the image data in the block based on the conversion table.
And fixed length coding means for coding each component into fixed length data.

According to the invention of claim 8, the block is subdivided into a plurality of small blocks for each of the C, M, Y, and K components, and the feature amount is extracted from the image data in the subdivided small blocks. Since the C, M, Y, and K data in each block are encoded into fixed-length data based on the extracted feature amount of the image data in each small block, even if the image is a color image, Without causing a decrease in coding efficiency,
It is possible to perform fixed-length coding in which partial characteristics of an image are reflected in coded data.

Further, by applying a low pass filter for obtaining the average of each pixel of the image data in the small block, C, M,
Applying a high-pass filter that extracts the pixel average of each of the Y and K components and obtains the difference between the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block, C,
Since the edge information of each of the M, Y, and K components is extracted, fixed-length coding can be performed in consideration of the edges of each of the C, M, Y, and K components.

Further, a conversion table in which the number of bits allocated to each small block and the conversion mode are associated with each other is provided, and the conversion mode is determined based on the presence or absence of edge information of each small block forming the block, and conversion is performed. C, M, Y, K forming image data in a block based on a table
Since each of the components is encoded into fixed length data, C, M, Y, K can be efficiently used by using the conversion table.
The components can be fixed length coded.

An image processing method according to a ninth aspect of the invention is an image processing method for dividing a multi-valued image into blocks of coding units and coding the image data in the divided blocks into fixed length data. A subdivision step of subdividing the block into a plurality of small blocks, an extraction step of extracting a feature amount from the image data in the small blocks subdivided by the subdivision step, and each extracted by the extraction step. And a coding step of coding the image data in the block into fixed-length data based on the feature amount of the image data in the small block, wherein the extracting step includes: A high-pass filter that finds the difference between the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block while applying a low-pass filter that calculates the average to extract the pixel average Filter to extract edge information, and the encoding step includes a conversion mode determination step of determining a conversion mode based on the presence or absence of edge information of each small block forming the block, and an allocation to each small block. And a fixed length encoding step of encoding the image data in the block into fixed length data based on a conversion table in which the number of bits and the conversion mode are associated with each other.

According to the invention of claim 9, the block is subdivided into a plurality of small blocks, the feature amount is extracted from the image data in the subdivided small blocks, and the image data in each extracted small block is extracted. Based on the feature amount, we decided to encode the image data in the block into fixed length data,
It is possible to perform fixed-length coding that reflects the partial characteristics of an image in code data without causing a reduction in coding efficiency.

A pixel average is extracted by applying a low-pass filter that calculates the average of each pixel of the image data in the small block, and the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block are extracted. Since we decided to apply a high-pass filter that calculates the difference of
Fixed-length coding that considers edges such as contour lines can be performed.

Further, a conversion table in which the number of bits allocated to each small block and the conversion mode are associated with each other is provided, and the conversion mode is determined based on the presence / absence of edge information of each small block forming the block. Since the image data in the block is encoded into the fixed length data based on the table, it is possible to efficiently perform the fixed length encoding using the conversion table.

The image processing method according to a tenth aspect of the invention is the image processing method according to the ninth aspect, wherein the conversion table is
It is characterized in that the number of bits assigned to the small block having no edge information and the number of bits assigned to the small block having the edge information are stored for each conversion mode.

According to the tenth aspect of the invention, the number of bits assigned to the small block having no edge information and the number of bits assigned to the small block having edge information are stored in the conversion table for each conversion mode. It is possible to assign a different number of bits depending on the presence or absence of an edge, and thus perform fixed length coding in consideration of the edge.

The image processing method according to an eleventh aspect of the present invention is the image processing method according to the tenth aspect of the invention, wherein the number of bits to be assigned to the small blocks not having the edge information in each conversion mode and the small blocks having the edge information are assigned. The sum with the number of bits is characterized by having a predetermined bit length.

According to the eleventh aspect of the present invention, the sum of the number of bits assigned to the small block having no edge information and the number of bits assigned to the small block having edge information in each conversion mode has a predetermined bit length. Therefore, the code data can always have a fixed length.

Further, the image processing method according to the invention of claim 12 is the image processing method according to claim 10 or 11, wherein, in the conversion table, the pixel position of a pixel forming an edge when the small block has edge information. The number of bits allocated to pixels forming the edge and the number of bits allocated to pixels not forming the edge are stored.

According to the twelfth aspect of the invention, when the small block has edge information, the pixel position of the pixel forming the edge, the number of bits allocated to the pixel forming the edge, and the pixel not forming the edge are allocated. Since the number of bits is stored, it is possible to perform fixed-length coding that reflects the pixel positions of the pixels forming the edge in the small block.

The image processing method according to a thirteenth aspect of the present invention is the image processing method according to the tenth aspect, the eleventh aspect, or the twelveth aspect of the invention, wherein in the fixed length encoding step, when the small block has edge information, the small block The bit string forming the pixel value of each pixel is truncated to the number of bits allocated to the pixels forming the edge or the bit string allocated to the pixels not forming the edge.

According to the thirteenth aspect of the invention, when the small block has edge information, the number of bits assigned to the pixels forming the edge or the bit string forming the pixel value of each pixel in the small block is not formed. Since the bit string assigned to the pixel is truncated, the data can be compressed quickly.

The image processing method according to claim 14 is the image processing method according to any one of claims 10 to 13, wherein the fixed length encoding step is performed by the extracting step when the small block has no edge information. It is characterized in that the image data in the small block is cut into a predetermined bit string based on the average of the extracted pixels.

According to the fourteenth aspect of the present invention, when the small block does not have edge information, the image data in the small block is truncated to a predetermined bit string based on the extracted pixel average. It is possible to efficiently perform data compression of pixels that do not form edges.

An image processing method according to a fifteenth aspect of the present invention is an image processing method for dividing a multi-valued image into blocks of a coding unit and coding the image data in the divided blocks into fixed length data. A multiplex division step of recursively multiplexing the block into small blocks having a hierarchical structure, and an extraction step of extracting edge information from the image data in the small block recursively multiplexed by the multiplex division step, An encoding step of encoding the image data in the block into fixed-length data based on the feature amount of the image data in each small block extracted by the extracting step, wherein the extracting step comprises: The pixel average is extracted by applying a low-pass filter that obtains the average of each pixel of the image data in the The edge information is extracted by applying a high-pass filter that obtains the difference between the prime value and the minimum pixel value, and the encoding step determines the conversion mode based on the presence or absence of the edge information of each small block forming the block. A conversion mode determining step, and a fixed length encoding step of encoding the image data in the block into fixed length data based on a conversion table in which the number of bits allocated to each small block and the conversion mode are associated with each other, It is characterized by including.

According to the fifteenth aspect of the present invention, while the block is multiplexed into small blocks, edge information is extracted from the image data in the small blocks, and the small blocks having the edge information are recursively re-divided. After repeating the process, based on the feature amount of the image data in the small block,
Since the image data in the block is encoded into the fixed length data, it is possible to improve the image quality by accurately reflecting the partial characteristics of the image in the code data.

Further, a low-pass filter for obtaining the average of each pixel of the image data in the small block is applied to extract the pixel average, and the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block are extracted. Since we decided to apply a high-pass filter that calculates the difference of
Fixed-length coding that considers edges such as contour lines can be performed.

Further, a conversion table in which the number of bits assigned to each small block and the conversion mode are associated with each other is provided, and the conversion mode is determined based on the presence / absence of edge information of each small block forming the block. Since the image data in the block is encoded into the fixed length data based on the table, it is possible to efficiently perform the fixed length encoding using the conversion table.

A recording medium according to a sixteenth aspect of the present invention records a program for causing a computer to execute the method according to the ninth aspect, so that the program can be machine-readable, whereby the recording medium according to the ninth aspect. The operation of can be realized by a computer.

[0052]

BEST MODE FOR CARRYING OUT THE INVENTION Preferred embodiments of an image processing apparatus, an image processing method, and a computer-readable recording medium recording a program for causing a computer to execute the method will be described below with reference to the accompanying drawings. The form will be described in detail.

In the present embodiment, the case where the present invention is applied to a facsimile apparatus will be described. Specifically, in the first embodiment, a wavelet (wavelet) of (2, 2) is used.
t) A case of extracting a feature amount using a transformation is shown, a case of extracting a feature amount using a differential operator and an integral operator is shown in the second embodiment, and a case is shown in the third embodiment.
A case where the present invention is applied to a color image will be shown.

(First Embodiment) FIG. 1 is a functional block diagram showing a configuration of a facsimile system according to the first embodiment. The facsimile system shown in FIG. 1 includes a transmission device 10 for transmitting a multivalued black and white grayscale image by fixed length coding for each block, and a transmission device 1.
The receiving device 11 decodes the coded data received from 0 for each block and outputs an image.

Specifically, the transmission device 10 includes an image reading unit 101, a block division processing unit 102, a block re-division processing unit 103, an image conversion processing unit 104, a conversion table 105, and an encoding processing unit. 106 and a data transmission unit 107.

The image reading section 101 optically reads a document placed on a document table (not shown) with a CCD sensor,
It is an input device that outputs the read image data to the block division processing unit 102. In the first embodiment, the image reading unit 101 has 256 pixels in the vertical and horizontal directions.
A monochrome grayscale image of pixels is input, and the pixel value of each pixel is 256 gradations.

The block division processing unit 102 is a processing unit that divides the image data read by the image reading unit 101 into rectangular blocks of a predetermined size. Specifically, such a block is a block having a size of an encoding processing unit,
Here, it is assumed that the size is 4 pixels × 4 pixels.

The block subdivision processing unit 103 is a processing unit that subdivides each block divided by the block division processing unit 102 into blocks (hereinafter referred to as “small blocks”) of a processing unit of the image conversion processing unit 104. . Here, the size of this small block is 2 pixels × 2 pixels.

The image conversion processing unit 104 is a processing unit for performing frequency conversion or the like on the image data in the small block re-divided by the block re-division processing unit 103. The feature quantity is extracted by performing the wavelet transform for extracting the edge.

The encoding processing unit 106 uses the conversion table 10
4 divided by the block division processing unit 102 using 5
This is a process of encoding the pixel value of each block of No. 4 into fixed length data. Specifically, instead of simply assigning the same number of bits to the pixel value of each pixel, a large number of bits are assigned to the pixels forming the edge portion while preserving the edge position so that the entire block has a fixed length. Encode.
Note that description of specific processing of the conversion table 105 and the encoding processing 106 will be given later.

The data transmission section 107 includes an encoding processing section 10
6 is a processing unit for transmitting the fixed length data encoded in 6 to the receiving device. The transmitting device 10 also has a display unit, an operation panel, and the like (not shown), but the description thereof is omitted here.

By using the transmitting device 10 having the above configuration, the image data read by the image reading unit 101 can be divided into blocks, and the image data in the blocks can be sequentially encoded into fixed-length data. . Here, in the present invention, since a large number of bits are assigned to the edge part of the image data, it is possible to reflect the partial characteristics of the image in the code data without lowering the coding efficiency. become.

Next, the configuration of the receiving apparatus 11 shown in FIG. 1 will be described. As shown in FIG.
Has a data receiving unit 111, a conversion table 112, a decoding processing unit 113, a block synthesis processing unit 114, an image processing unit 115, and an image output unit 116.

The data receiving unit 111 is a processing unit for sequentially receiving the fixed length data transmitted from the transmitting device 10 via the line. The conversion table 112 is a table that stores the same contents as the conversion table 105 in the transmission device 10.

The decoding processing unit 113 has a conversion table 112.
Is a processing unit that decodes the fixed-length data received by the data receiving unit 111 into pixel values of a 4 × 4 block by using. In addition, in the decoding processing unit 113, when it is set in advance that random noise is added after the decoding processing of the image data is finished using the conversion table 112, the pseudo random number series such as the M series is used. Accordingly, a random number is sequentially generated and a process of embedding it in the image data is performed. The reason for embedding such noise is to reduce the block distortion caused by the density difference at the block boundary and the same pixel value within the block, and to reproduce a higher quality image.

The block synthesis processing section 114 is a processing section for sequentially synthesizing the image data of each block obtained by decoding the fixed length data by the decoding processing section 113, and finally forming the entire image data.

The image processing section 115 is a processing section for performing various kinds of image processing on the combined image data. The image output unit 116 is a processing unit that outputs image data that has been subjected to image processing by the image processing unit 115. Specifically, the image output unit 116 displays the image data on a display unit (not shown), prints the image data on printing paper, and the like. Perform.

By using the receiving device 11 having the above-mentioned structure, the partial characteristics of the image are not impaired,
It is possible to efficiently restore fixed-length encoded code data.

Next, the block division processing unit 1 shown in FIG.
02 and block division by the block re-division processing unit 103 will be described more specifically. FIG. 2 is an explanatory diagram for explaining the concept of block division by the block division processing unit 102 and the block re-division processing unit 103 shown in FIG. For convenience of explanation, the size of the image data is 8 × 8.

As shown in FIG. 9A, the block division processing unit 102 is 8 × 8 read by the image reading unit 101.
Image data of four 4 × 4 blocks a0, a1, a
2, a3. The block size is set to 4 in order to match the coding processing unit by the coding processing unit 106.
It is set to x4. The block division processing unit 102
Is not unique to the present invention and is widely adopted in various image processing apparatuses that perform block coding.

On the other hand, the block re-division processing unit 10
3 is related to the present invention, and specifically, FIG.
, The blocks divided by the block division processing unit 102 are divided into 2 × 2 small blocks b0, b1, b2.
It is further divided into b3. The reason for dividing into 2 × 2 small blocks is to match the processing unit of the (2,2) wavelet conversion used in the image conversion processing unit 104, and the reason for performing this wavelet conversion is
This is to realize fixed-length coding that preserves edges.

Here, for convenience of explanation, the case where the image data is divided into 4 × 4 blocks and then subdivided into 2 × 2 small blocks is shown.
It is also possible to divide it into m (n, m is an integer) rectangular block, and further divide this into (n / p) × (m / r) (p, r are integers) small blocks. In this case, the encoding process is performed in units of n × m rectangular blocks, and (n / p)
Image conversion is performed in small block units of x (m / r).

Next, the image conversion processing unit 104 shown in FIG.
The frequency conversion processing performed in step 2 will be specifically described.
3 and 4 show the image conversion processing unit 104 shown in FIG.
It is explanatory drawing for demonstrating the concept of the frequency conversion process performed by.

As shown in FIG. 3, in the image conversion processing unit 104, the high pass filter 104a and the low pass filter 104b are applied to the image data in the small block. For example, this high pass filter 1
As an example of 04a, there is a method of obtaining the difference between the maximum value and the minimum value of the pixels in the small block in consideration of the direction. As an example of the low pass filter 104b, the average of the pixels in the small block is obtained. I can give you something.

As a technique for executing the high pass filter 104a and the low pass filter 104b in one pass,
The wavelet transform shown in FIG. 4 is known. As shown in the figure, when the pixel values of four pixels in a 2 × 2 small block are a, b, c, and d, respectively, the low-pass filter (L) is

The processing of L = (a + b + c + d) / 4 is performed, and the high pass filters (H1, H2, H
3) is

[0077] H1 = (a + c) / 2- (b + d) / 2

[0078] H2 = (a + b) / 2- (c + d) / 2

[0079] H3 = (ab)-(cd) Will be processed.

Next, the fixed length coding process performed by the coding processing unit 106 shown in FIG. 1 will be described in detail.
FIG. 5 is a diagram showing an example of the conversion table 105 shown in FIG. 1, and FIG. 6 is an explanatory diagram for explaining modes 0 to 15 determined based on the presence / absence of edges of small blocks.

First, as shown in FIG. 5A, the pixel values of the pixels in the block are set to d0 to d15, and as shown in FIG. 5B, the four small blocks forming the block are b0. , B1, b2, b3.

When it is confirmed whether or not each pixel in the small block forms a part of the edge, each block coincides with any one of the modes 0 to 15, as shown in FIG. Specifically, this mode 0 shows the case where no edge exists in any of the small blocks, and the mode 1 shows the case where the edge exists only in the upper left small block b0. Similarly, mode 15 shows a case where an edge exists in any of the small blocks b0 to b3.

Then, in the conversion table 105, FIG.
As shown in (c), the number of bits assigned to each small block is defined for each mode. This conversion table 105
The reason for using is to perform different encoding for each mode. In other words, if different encoding is performed for each mode, the encoding method is changed depending on whether or not an edge exists, and fixed length encoding can be performed while preserving the edge.

Specifically, in the case of mode 0, 7 bits are assigned to each of the small blocks b0, b1, b2 and b3. Since there are 4 mode bits indicating the type of mode, the total is 32 bits.

Here, the average of the pixel values obtained by applying the low-pass filter 104b to each small block (hereinafter referred to as "pixel average") is 7 bits for each small block. When the pixel average is 8 bits or more, 7 bits are selected from the upper bits and the lower bits are discarded to perform the truncation. The reason for encoding using the 7-bit pixel average in the case of mode 0 is that there is no edge in any small block in this case.

In mode 1, 5 bits are assigned to each of the small blocks b1, b2, and b3 having no edge, and 1 is assigned to the small block b0 having an edge.
Allocate 3 bits. Therefore, the total of the mode bits is 32 bits.

Here, the small block b0 where the edge exists
Is 4 for the pixel value (min) of the pixel that is not an edge.
5 bits are assigned to the pixel value (max) of the pixel forming the bit and the edge, and 4 bits are assigned to the edge position information indicating whether each pixel in the small block b0 forms the edge.

Therefore, in the case of this mode 1, 5 bits are allocated to the pixel forming the edge, and 4 bits (in the case of the small block b0) or 5 bits (the small block b1, to the pixel not forming the edge). b2, b3) will be allocated.

When 16 pixels in one block are fixed-length coded in 32 bits by the conventional fixed-length coding technique, each pixel is expressed in 2 bits regardless of whether an edge is formed or not. Therefore, it can be seen that the encoding of the present invention significantly improves the resolution.

Similarly, in the mode 15, the pixel value (min) of a pixel that is not an edge is 6 bits, the pixel value (max) of a pixel that is an edge is 6 bits, and the small block b0,
16 bits are allocated to edge position information indicating whether or not each pixel in b1, b2, and b3 forms an edge. Therefore, the total number of bits together with the mode bits is 32 bits.
In addition, in this mode 15, since all the small blocks include pixels forming edges, the pixel average is not used.

As described above, by using the conversion table 105, the encoding processing unit 106 can change the encoding method depending on whether or not there is a pixel forming an edge in each small block, and the edge is saved. It is possible to perform the fixed length coding described above. The conversion table 112 included in the reception device 11 is also the same as the conversion table 105, and the decoding processing unit 113 performs the decoding process using the conversion table 112.

Next, the decoding processing unit 113 shown in FIG. 1 will be described more specifically. FIG. 7 is a functional block diagram showing a detailed configuration of the decoding processing unit 113 shown in FIG. As shown in the figure, the decoding processing unit 113 includes a decoding unit 113a, a switching unit 113b, and a noise addition processing unit 1.
13c.

The decoding section 113a is a processing section for decoding the image data of each block from the fixed length data based on the conversion table 112. Specifically, when the conversion table 112 shown in FIG. 5C is included, the image data is decoded according to the mode bit of the fixed length data.

For example, when the mode bit of the fixed length data indicates mode 0, bits 0 to 6 are the pixel values of the respective pixels in the small block b3, and bits 7 to 13 are in the small block b2. Let the pixel value of each pixel be the bit 14, the bit 20 be the pixel value of each pixel in the small block b1, and the bit 21 be the bit 27 be the pixel value of each pixel in the small block b0.

The switching unit 113b is a switching switch for switching whether or not to add noise to the image data decoded by the decoding unit 113a, and when noise addition is designated from the operation panel or the like, The image data is output to the noise addition processing unit 113c, and when the designation is not made, the image data is directly output to the block synthesis processing unit 114 instead of the noise addition processing unit 113c.

The noise addition processing unit 113c is a processing unit that sequentially generates a pseudorandom number sequence (M sequence) having a maximum long cycle based on a predetermined primitive irreducible polynomial and adds the generated pseudorandom number to the image data as noise. is there. Specifically, FIG.
In the mode 0 of (c), since the pixel value of each pixel is truncated to 7 bits, noise is added to the least significant bit. Further, in the case of mode 1, for example, since the pixel value of each pixel of the small block b1 is truncated to 5 bits, noise is added to the lower 3 bits.

As described above, when noise is added to the image data in the block, the block distortion caused by the difference in density at the block boundary and the same pixel value in the block is reduced, so that a high-quality image is obtained. Can be reproduced.

Next, the processing procedure of the transmitting apparatus 10 shown in FIG. 1 will be described. FIG. 8 shows the transmitter 1 shown in FIG.
It is a flow chart which shows the processing procedure of 0. As shown in the figure, in the transmitting device 10, if the image data is read by the image reading unit 101 (step S80).
1) The block division processing unit 102 divides the image data into blocks that are units of encoding processing (step S8).
02).

Thereafter, the block re-division processing unit 103
Each of the divided blocks is divided into small blocks as a processing unit of the image conversion processing unit 104 (step S803),
The image conversion processing unit 104 performs wavelet conversion on each small block to obtain edge and pixel averages (steps S804 to S805).

Then, the encoding processing unit 106 selects a mode based on the presence / absence of an edge in the small block (step S806), and encodes the image data in the block into fixed-length data according to the selected mode ((S806)). In step S807), the encoded fixed length data is transmitted (step S808).

Next, the processing procedure of the receiving apparatus 11 shown in FIG. 1 will be described. FIG. 9 shows the receiving device 1 shown in FIG.
It is a flow chart which shows a processing procedure of 1. As shown in the figure, in the receiving device 11, if the data receiving unit 111 sequentially receives each fixed length data transmitted from the transmitting device 10 via the line (step S901), the decoding processing unit 113 determines The pixel value in the block is restored according to the mode bit included in the fixed length data and the conversion table 112 (step S902).

Then, it is confirmed whether or not the setting for adding random noise is set (step S90).
3) If noise is set to be added (Yes at step S903), random noise is generated and added to the pixel value of each pixel (step S904).

Then, the decoded blocks are sequentially combined to generate image data (step S905), and the generated image data is displayed on the display unit or printed on a printing paper (step S906).

Next, the conversion table 1 shown in FIG.
The advantage of performing fixed-length coding using 05 will be described. As described above, the conversion table 10
Performing fixed-length coding using 5 has the advantage that partial characteristics of an image can be reflected in code data without degrading coding efficiency, and rotation in the code state can be performed.
It has the advantage that it can be mirrored or edited.

In FIG. 10, 90 degrees clockwise in the code state.
FIG. 12 is an explanatory diagram for explaining a case of rotating by one degree, and FIG.
[Fig. 3] is an explanatory diagram for explaining a case where mirroring is performed in a coded state.

As shown in FIG. 10, the case of rotating 90 degrees can be dealt with by changing the mode of the conversion table 105 and changing the bit order. Specifically, the mode 1 is changed to the mode 2 and the mode 2 is changed to the mode 8
By changing the mode bits such as changing the mode 3 to the mode 10 and changing the order of the bits for each mode, the image can be easily rotated 90 degrees in the code state.

As shown in FIG. 11, mirroring can also be handled by changing the mode of the conversion table 105 and changing the bit order. Specifically, mode 1 is changed to mode 2, mode 2 is changed to mode 1, mode 4 is changed to mode 8, and the mode bits are changed, and the bit order of each mode is changed to change the code. You can easily mirror the image as it is.

Even when the image is rotated by 180 degrees or 270 degrees, the image can be rotated in the coded state by simply changing the mode and changing the bit order.

As described above, in the first embodiment, the block subdivision processing unit 103 subdivides the block into a plurality of small blocks, and the image conversion processing unit 104 extracts edges and pixels from the image data in the small blocks. Extract the average,
Since the encoding processing unit 106 is configured to encode the image data in the block into the fixed length data based on the presence / absence of the edge in each extracted small block and the conversion table 105, the encoding efficiency is deteriorated. Instead, it is possible to perform fixed-length coding in which partial characteristics of an image are reflected in code data.

(Embodiment 2) By the way, in the above-mentioned Embodiment 1, the edge and the pixel average in the small block are obtained by using the wavelet transform of (2, 2). Edges and pixel averages can also be determined by applying filters respectively.

Therefore, in the second embodiment, the case where the edge and pixel averages are obtained by applying the differential filter and the local average filter separately will be described. In addition, here, it is assumed that the size of the small block is 3 × 3.

FIG. 12 is a functional block diagram showing the structure of the image conversion processing unit 201 used in the second embodiment. As shown in FIG. 7A, the image conversion processing unit 201
It has a differential filter 202 and a local average filter 203.

Specifically, as the differential filter 202, a Sobel operator shown in FIG. 9B can be used, and as the local average filter 203, the weight shown in FIG. A local weighted average filter using the coefficient matrix W can be used.

Although the case where the edge and the pixel average are used by using the local filter is shown here, the present invention is not limited to this, and Fourier transform or discrete cosine transform can be used.

(Embodiment 3) In the above Embodiments 1 and 2, the case where a block is subdivided into small blocks and encoded is shown, but such subdivision can also be performed in multiplex. Therefore, in the third embodiment, a case will be described in which the redivision is performed in multiplex. Note that here, the unit of encoding processing is 8 × 8.
A pixel is used, and the processing unit of the image conversion processing unit 104 is 2 × 2 pixels.

FIG. 13 is an explanatory diagram for explaining a case where the subdivision is recursively repeated. FIG. 11A shows a block (target image) divided into coding units (8 × 8), and pixels that contribute to the image quality as shown in FIG.

Here, when this block (target image) is not recursively subdivided, as shown in FIG.
This block is divided into 2 × 2 small blocks, and the blocks are coded based on the presence / absence of edges in the small blocks.

However, if such subdivision is performed, the pixel average of A1 shown in the figure cannot be used for a minute block having no edge, and the pixel average in the small block A0 cannot be used. , The encoding efficiency cannot be improved.

On the other hand, as shown in FIG.
When recursively subdividing a block (target image),
First, the block is divided into 4 × 4 small blocks, and then the small block having the characteristic pixel is recursively divided into 2 × 2 small blocks.

As described above, when the block is recursively subdivided, the pixel average of the 4 × 4 small block A0 can be used. Therefore, the total number of bits should be reduced as compared with the case shown in FIG. You will be able to Since the reduced number of bits can be allocated to characteristic pixels, the image quality can be improved as a result when recursive redivision is performed.

As described above, in the third embodiment, the block is recursively re-divided.
It is possible to reduce the number of bits to be assigned to non-characteristic pixels and to assign the bits to characteristic pixels, thereby improving the image quality. For convenience of explanation, the case where the subdivision is performed twice is shown here, but the present invention can be applied to the case where the subdivision is performed three times or more.

In the above-described first to third embodiments, the case where the present invention is applied to a monochrome grayscale image has been shown, but the present invention is not limited to this, and it is applicable to a multivalued color image. You can also

Specifically, in the case of a CMYK color system color image, the transmitting apparatus 10 applies the present invention to each of the C, M, Y and K components to obtain fixed length data of each component. , The acquired fixed length data respectively
Send to. On the other hand, the receiving device 11 applies the present invention to each of the C, M, Y, and K components to restore the image data of each component, synthesizes the restored image data, and outputs the synthesized image data.

[0124]

As described above, according to the invention of claim 1, the block is subdivided into a plurality of small blocks, and the feature amount is extracted from the image data in the subdivided small blocks.
Since the image data in the block is encoded into the fixed length data based on the extracted feature amount of the image data in each small block, the partial efficiency of the image can be maintained without lowering the encoding efficiency. An effect is obtained that an image processing device capable of performing fixed-length coding that reflects characteristics in code data is obtained.

Further, a conversion table in which the number of bits allocated to each small block and the conversion mode are associated with each other is provided, and the conversion mode is determined based on the presence / absence of edge information of each small block forming the block. Since the image data in the block is encoded into the fixed length data based on the table, it is possible to obtain the image processing device capable of efficiently performing the fixed length encoding by using the conversion table. .

According to the second aspect of the invention, the number of bits assigned to the small blocks having no edge information and the number of bits assigned to the small blocks having edge information are stored in the conversion table for each conversion mode. Therefore, there is an effect that a different number of bits is assigned depending on the presence or absence of an edge, and thus an image processing apparatus capable of performing fixed-length coding in consideration of the edge can be obtained.

According to the third aspect of the present invention, the sum of the number of bits assigned to the small blocks having no edge information and the number of bits assigned to the small blocks having edge information in each conversion mode has a predetermined bit length. With this configuration, it is possible to obtain an image processing device that can always have a fixed length of code data.

According to the invention of claim 4, when the small block has edge information, the pixel position of the pixel forming the edge, the number of bits allocated to the pixel forming the edge, and the pixel not forming the edge are set. Since the number of bits to be allocated is stored, it is possible to obtain an image processing device capable of performing fixed-length coding that reflects the pixel positions of pixels forming an edge in a small block.

According to the fifth aspect of the invention, when the small block has edge information, the number of bits to be assigned to the pixels forming the edge or the bit string forming the pixel value of each pixel in the small block is formed. Since the bit string to be assigned to the non-pixels is truncated, the image processing device capable of quickly compressing the data can be obtained.

Further, according to the invention of claim 6, when the small block does not have edge information, the image data in the small block is truncated to a predetermined bit string based on the average of the extracted pixels. Therefore, it is possible to obtain an image processing apparatus that can efficiently perform data compression of pixels that do not form edges.

Further, according to the invention of claim 7, while dividing the block into small blocks, edge information is extracted from the image data in the small blocks, and the small blocks having the edge information are recursively reproduced. After repeating the dividing process, based on the feature amount of the image data in the small block,
Since the image data in the block is configured to be encoded into the fixed length data, the image processing device capable of accurately reflecting the partial characteristics of the image in the code data and improving the image quality is provided. The effect is obtained.

Further, a low-pass filter for obtaining the average of each pixel of the image data in the small block is applied to extract the pixel average, and the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block are extracted. Since the configuration is such that the edge information is extracted by applying the high-pass filter that obtains the difference between the two, it is possible to obtain an image processing device capable of performing fixed-length encoding in consideration of edges such as contour lines. .

Further, a conversion table in which the number of bits allocated to each small block and the conversion mode are associated with each other is provided, and the conversion mode is determined based on the presence / absence of edge information of each small block forming the block, and conversion is performed. Since the image data in the block is encoded into the fixed length data based on the table, it is possible to obtain the image processing device capable of efficiently performing the fixed length encoding by using the conversion table. .

According to the invention of claim 8, the block is subdivided into a plurality of small blocks for each of the C, M, Y, and K components, and the feature amount is calculated from the image data in the subdivided small blocks. Since it is configured to encode the C, M, Y, and K data in each block into fixed-length data based on the extracted feature amount of the image data in each small block,
Even in the case of a color image, it is possible to obtain an image processing apparatus capable of performing fixed-length coding that reflects the partial characteristics of the image in the code data without causing a decrease in coding efficiency. .

Further, by applying a low pass filter for obtaining the average of each pixel of the image data in the small block, C, M,
Applying a high-pass filter that extracts the pixel average of each of the Y and K components and obtains the difference between the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block, C,
Since the edge information of each of the M, Y, and K components is configured to be extracted, it is possible to obtain an image processing apparatus capable of performing fixed-length coding in consideration of the edges of each of the C, M, Y, and K components. Play.

Further, a conversion table in which the number of bits allocated to each small block and the conversion mode are associated with each other is provided, and the conversion mode is determined based on the presence / absence of edge information of each small block forming the block. C, M, Y, K forming image data in a block based on a table
Since each component is configured to be encoded into fixed length data, the conversion table is used to efficiently perform C, M, Y, and
There is an effect that an image processing method capable of performing fixed-length coding on the K component can be obtained.

Further, according to the invention of claim 9, the block is subdivided into a plurality of small blocks, the feature amount is extracted from the image data in the subdivided small blocks, and the image data in each extracted small block is extracted. Since the image data in the block is encoded into fixed-length data based on the feature amount of, the fixed-length data that reflects the partial feature of the image in the code data without lowering the encoding efficiency An effect that an image processing method capable of encoding can be obtained.

Further, a low pass filter for obtaining the average of each pixel of the image data in the small block is applied to extract the pixel average, and the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block are extracted. Since it is configured to extract edge information by applying a high-pass filter that obtains the difference between the two, it is possible to obtain an image processing method capable of performing fixed-length coding in consideration of edges such as contour lines. .

Further, a conversion table in which the number of bits allocated to each small block and the conversion mode are associated with each other is provided, and the conversion mode is determined based on the presence / absence of edge information of each small block forming the block. Since the image data in the block is encoded into the fixed length data based on the table, it is possible to obtain the image processing method capable of efficiently performing the fixed length encoding by using the conversion table. .

According to the tenth aspect of the invention, the number of bits assigned to the small blocks having no edge information and the number of bits assigned to the small blocks having edge information are stored in the conversion table for each conversion mode. Therefore, there is an effect that a different number of bits is assigned depending on the presence / absence of an edge, and thus an image processing method capable of performing fixed-length coding in consideration of the edge can be obtained.

According to the eleventh aspect of the present invention, the sum of the number of bits assigned to the small blocks having no edge information and the number of bits assigned to the small blocks having edge information in each conversion mode has a predetermined bit length. With this configuration, it is possible to obtain an image processing method in which code data can always have a fixed length.

According to the invention of claim 12, when the small block has edge information, the pixel position of the pixel forming the edge, the number of bits to be allocated to the pixel forming the edge, and the pixel not forming the edge are set. Since the number of bits to be allocated is stored respectively, there is an effect that it is possible to obtain an image processing method capable of performing fixed-length coding that reflects the pixel position of a pixel forming an edge in a small block.

According to the thirteenth aspect of the invention, when the small block has edge information, the number of bits to be assigned to the pixel forming the edge or the edge is formed by forming the bit string forming the pixel value of each pixel in the small block. Since the bit string to be assigned to the non-pixels is truncated, an image processing method capable of quickly compressing the data can be obtained.

According to the fourteenth aspect of the invention, when the small block has no edge information, the image data in the small block is truncated to a predetermined bit string based on the extracted pixel average. Thus, there is an effect that an image processing method capable of efficiently performing data compression of pixels that do not form edges is obtained.

According to the fifteenth aspect of the invention, while dividing the block into small blocks, the edge information is extracted from the image data in the small blocks, and the small blocks having the edge information are recursively reproduced. After repeating the dividing process, the image data in the block is configured to be encoded into the fixed length data based on the feature amount of the image data in the small block. An effect is obtained that an image processing method capable of accurately reflecting in code data and improving image quality can be obtained.

Further, a low-pass filter for obtaining the average of each pixel of the image data in the small block is applied to extract the pixel average, and the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block are extracted. Since it is configured to extract edge information by applying a high-pass filter that obtains the difference of, there is an effect that an image processing method capable of performing fixed-length coding in consideration of edges such as contour lines can be obtained. .

Further, a conversion table in which the number of bits allocated to each small block and the conversion mode are associated with each other is provided, and the conversion mode is determined based on the presence / absence of edge information of each small block forming the block. Since the image data in the block is encoded into the fixed length data based on the table, it is possible to obtain the image processing method capable of efficiently performing the fixed length encoding by using the conversion table. .

According to the sixteenth aspect of the present invention, by recording the program for causing a computer to execute the method according to the ninth aspect, the program becomes machine-readable, whereby the operation of the ninth aspect is performed. There is an effect that a recording medium that can be realized by a computer can be obtained.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing a configuration of a facsimile system according to a first embodiment of the present invention.

FIG. 2 is an explanatory diagram for explaining the concept of block division by a block division processing unit and a block re-division processing unit shown in FIG.

FIG. 3 is an explanatory diagram for explaining the concept of frequency conversion processing performed by the image conversion processing unit shown in FIG. 1.

FIG. 4 is an explanatory diagram for explaining the concept of frequency conversion processing performed by the image conversion processing unit shown in FIG. 1.

5 is a diagram showing an example of a conversion table shown in FIG.

FIG. 6 is an explanatory diagram for explaining modes 0 to 15 that are determined based on the presence / absence of edges of small blocks.

7 is a functional block diagram showing a detailed configuration of a decoding processing unit shown in FIG.

8 is a flowchart showing a processing procedure of the transmission device shown in FIG.

9 is a flowchart showing a processing procedure of the receiving apparatus shown in FIG.

FIG. 10 is an explanatory diagram for explaining a case of rotating clockwise by 90 degrees in a coded state.

FIG. 11 is an explanatory diagram for explaining a case where mirroring is performed in the code state.

FIG. 12 is a functional block diagram showing a configuration of an image conversion processing unit used in the second embodiment.

FIG. 13 is an explanatory diagram for explaining a case where re-division is recursively repeated.

[Explanation of symbols]

10 transmitter 11 Receiver 101 image reading unit 102 block division processing unit 103 block re-division processing unit 104 image conversion processing unit 105 conversion table 106 encoding processing unit 107 data transmitter 111 Data receiver 112 conversion table 113 decryption processing unit 114 block synthesis processing unit 115 Image processing unit 116 Image output unit

Continuation of front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) H04N 1/415

Claims (16)

(57) [Claims] 1. An image processing apparatus that divides a multi-valued image into blocks of coding units and encodes image data in the divided blocks into fixed-length data, wherein the blocks are subdivided into a plurality of small blocks. Based on the feature amount of the image data in each small block extracted by the extracting unit, a re-dividing unit, an extracting unit for extracting a feature amount from the image data in the small block redivided by the re-dividing unit, , Determining the number of bits allocated to each small block, and encoding means for encoding the image data in the block into fixed-length data, the extraction means, each of the pixels of the image data in the small block A high-pass for obtaining the difference between the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block while applying a low-pass filter for obtaining the average to extract the pixel average The edge information is extracted by applying an overfilter, the encoding means determines a conversion mode based on the presence or absence of edge information of each small block forming the block, and a conversion mode determining means for each small block. A conversion table in which the number of allocated bits and the conversion mode are associated with each other,
An image processing device, comprising: a fixed-length encoding unit that encodes image data in the block into fixed-length data based on the conversion table. 2. The conversion table stores, for each conversion mode, the number of bits to be allocated to the small block having no edge information and the number of bits to be allocated to the small block having the edge information. The image processing device according to item 1. 3. The sum of the number of bits assigned to the small block having no edge information and the number of bits assigned to the small block having the edge information in each conversion mode has a predetermined bit length. Item 2. The image processing device according to item 2. 4. The conversion table, when the small block has edge information, the pixel position of a pixel forming an edge, the number of bits allocated to the pixel forming the edge, and the bit allocated to the pixel not forming the edge. 4. The number according to claim 2, wherein the numbers are stored respectively.
The image processing device according to item 1. 5. The fixed length encoding means, when the small block has edge information, the number of bits assigned to the pixels forming the edge or the bit number forming the pixel value of each pixel in the small block, or 5. The image processing device according to claim 2, wherein the bit string is assigned to a pixel that does not form an edge. 6. The fixed length coding means truncates the image data in the small block into a predetermined bit string based on the pixel average extracted by the extracting means when the small block does not have edge information. The image processing apparatus according to any one of claims 2 to 5, characterized in that. 7. An image processing apparatus for dividing a multi-valued image into blocks of coding units and coding image data in the divided blocks into fixed length data, wherein the blocks are recursed into small blocks having a hierarchical structure. Of the image data in each small block extracted by the extracting unit, and an extracting unit that extracts edge information from the image data in the small block that is multiplexed by the multiple dividing unit. Encoding means for encoding the image data in the block into fixed-length data based on the characteristic amount, and the extracting means obtains an average of each pixel of the image data in the small block. A high-pass filter for extracting the pixel average by applying Applying to extract edge information, the encoding means determines a conversion mode based on the presence / absence of edge information of each small block forming the block, and the number of bits allocated to each small block. And a conversion table in which the conversion mode is associated with each other, and fixed-length coding means for coding the image data in the block into fixed-length data based on the conversion table. apparatus. 8. An image processing apparatus for dividing a color image of CMYK color system into blocks of encoding units and encoding image data in the divided blocks into fixed length data, wherein the blocks are C, M, A subdivision unit that subdivides each of Y and K components into a plurality of small blocks, an extraction unit that extracts a feature amount from the image data in the small blocks subdivided by the subdivision unit, and the extraction unit. Based on the feature amount of the extracted image data in each small block, C, M, and
Coding means for respectively coding Y and K data into fixed-length data, wherein the extracting means applies a low-pass filter for obtaining an average of each pixel of the image data in the small block to C, M ,
C, M, Y and K are applied by extracting the pixel average of each of the Y and K components and applying a high-pass filter that finds the difference between the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block. Extracting edge information of each component, the encoding means determines a conversion mode based on the presence / absence of edge information of each small block forming the block, and the number of bits allocated to each small block. And a conversion table in which the conversion modes are associated with each other,
An image processing apparatus, comprising: fixed-length encoding means for encoding each of C, M, Y, and K components forming image data in the block into fixed-length data based on the conversion table. . 9. An image processing method for dividing a multi-valued image into blocks of a coding unit and coding image data in the divided blocks into fixed length data, wherein the block is subdivided into a plurality of small blocks. A redivision step, an extraction step of extracting a feature amount from the image data in the small block redivided by the redivision step, and a feature amount of the image data in each small block extracted by the extraction step And an encoding step of encoding the image data in the block into fixed-length data, the extracting step applying a low-pass filter to obtain an average of each pixel of the image data in the small block While extracting the average, the edge information is extracted by applying a high-pass filter that obtains the difference between the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block, The encoding step associates the conversion mode determining step of determining the conversion mode based on the presence or absence of edge information of each small block forming the block, the number of bits allocated to each small block, and the conversion mode. A fixed length encoding step of encoding the image data in the block into fixed length data based on a conversion table. 10. The conversion table stores, for each conversion mode, the number of bits to be allocated to the small block having no edge information and the number of bits to be allocated to the small block having the edge information. The image processing method described in. 11. The sum of the number of bits assigned to the small block not having the edge information and the number of bits assigned to the small block having the edge information in each conversion mode has a predetermined bit length. Item 10
The image processing method described in. 12. The conversion table, when the small block has edge information, the pixel position of a pixel forming an edge, the number of bits allocated to the pixel forming the edge, and the bit allocated to the pixel not forming the edge. The image processing method according to claim 10, wherein the numbers are stored respectively. 13. The fixed length encoding step, wherein when the small block has edge information, the number of bits assigned to a pixel forming the edge or a bit string forming a pixel value of each pixel in the small block, or 13. The image processing method according to claim 10, wherein the bit string is assigned to a pixel that does not form an edge. 14. The fixed length encoding step truncates image data in the small block into a predetermined bit string based on the pixel average extracted in the extracting step when the small block does not have edge information. The image processing method according to any one of claims 10 to 13, characterized in that. 15. An image processing method for dividing a multi-valued image into blocks of a coding unit and coding image data in the divided blocks into fixed-length data, wherein the blocks are recursed into small blocks having a hierarchical structure. A multiplex division step of performing multiple divisions, an extraction step of extracting edge information from image data in the small blocks recursively divided in the multiple division step, and a sub-block of each small block extracted by the extraction step. A coding step of coding the image data in the block into fixed-length data based on the feature amount of the image data, wherein the extracting step calculates the average of each pixel of the image data in the small block. A high-pass filter that applies a band-pass filter to extract the pixel average and obtains the difference between the maximum pixel value and the minimum pixel value of each pixel of the image data in the small block. Applying a filter to extract edge information, the encoding step includes a conversion mode determination step of determining a conversion mode based on the presence or absence of edge information of each small block forming the block, and an allocation to each small block. An image processing method, comprising: a fixed-length encoding step of encoding image data in the block into fixed-length data based on a conversion table in which the number of bits and the conversion mode are associated with each other. 16. A computer-readable recording medium in which a program for causing a computer to execute the method according to claim 9 is recorded.