JPH11220728A - Context generation method at image compression, image compression method, and image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a context generation method, an image compression method and an image processor for generating reference state (or context) to improve the compression rate, without increasing the number of the context (the number of reference tables) so much. SOLUTION: As a bit arrangement of address data for deciding a context at calculation encoding processing in image compression, besides a surrounding image element value m bit of pixel of attention, values Q1,..., Qn other than a pixel value of this surrounding pixel are operated, and the operation result P1+...+Pn is added to the bit arrangement of the address data for context decision. In the operation, as a result of binarization by an average value of the pixel value of an input pixel, an average density and the result of binarization by a corresponding dither threshold, quality and attribute information of an image, including an edge part and a flat part of the image, a change rate of the average density in an area of plural pixels, the size and positive or negative signs and so on are calculated, and binary or a multi-valued output is set as the operation results.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a context generation method for generating a context to be referred to when compressing an image by arithmetic coding, and further relates to an image compression method using the context and an image processing apparatus to which the image compression method is applied. It is.

[0002]

2. Description of the Related Art Conventionally, as an information storage type compression technique by an image compression technique, an arithmetic coder is used because of its high efficiency, and is also adopted in a standard compression method "JBIG" for a binary image. In this arithmetic coder, one or several pixels around the pixel of interest to be compressed are used as reference pixels, and a context is generated using the arrangement of each bit as an address, and for each bit state of the reference pixel, By managing the appearance probability estimation value for estimating the target pixel as the content of the context, efficient entropy coding is performed.

[0003]

However, in the above conventional example, when the number of reference pixels is increased in order to increase the compression efficiency, the capacity of the RAM table for holding the contents of the context increases, and the burden on hardware increases. Become. or,
In the case of pixels that do not affect the probability estimation even if the number of reference pixels is simply increased, the occurrence frequency of each state decreases by the increase in the number of reference tables. It can have adverse effects.

The present invention provides a context generation method, an image compression method, and an image processing apparatus for generating a reference state (context) for improving the compression ratio without increasing the number of contexts (the number of reference tables) so much. With the goal.

[0005]

In order to solve this problem, a context generating method according to the present invention uses, as a bit array of address data for determining a context at the time of arithmetic code processing in image compression, a peripheral pixel of a target pixel. In addition to the value, a value other than the pixel value of the peripheral pixel is calculated, and the calculation result is added to the bit array of the address data for context determination.

Here, a plurality of different types of calculations are performed, and the calculation results are added to the bit array of the address data for context determination. Further, the value other than the pixel value of the peripheral pixel is a pixel value including peripheral pixels other than the peripheral pixel. In the calculation, the average value of the pixel values of the input pixels is obtained, and a binary or multi-value output is used as the calculation result. In addition, a result of binarization based on the average density and the corresponding dither threshold is output as the calculation result. In the calculation, the property and attribute information of the image including the edge portion and the flat portion of the image are output as binary or multi-valued data. In addition, in the calculation, the change rate, the magnitude, the magnitude and the sign of the average density in the area of a plurality of pixels are calculated and output. In the calculation, the coordinates of the pixel of interest are input, and as a result of the calculation, a dither threshold value is calculated from a value converted into a coordinate value on a dither matrix, binary or multi-valued data, or a coordinate value on the dither matrix. A dither threshold corresponding to an input or data obtained by binarizing or multi-leveling the dither threshold is output.

Further, according to the image compression method of the present invention, there is provided an image compression method for determining a context using a peripheral pixel value of a pixel of interest as a bit array of address data and performing arithmetic coding based on the context. As a bit array of address data that determines the context at the time of the encoding process, in addition to the peripheral pixel value of the target pixel, a value other than the pixel value of the peripheral pixel is calculated, and the calculation result is expressed as a bit of the address data of the context determination. It is characterized in that it is added to an array. Here, a plurality of different types of operations are performed, and the operation results are respectively added to the bit array of the address data for context determination.

Further, the image compression apparatus of the present invention has context determination means for determining a context by using a peripheral pixel value of a target pixel as a bit array of address data, and performs an arithmetic coding based on the context. In the above, the context determination means calculates a value other than the pixel value of the peripheral pixel in addition to the peripheral pixel value of the target pixel as a bit array of address data for determining the context in arithmetic code processing in image compression. And an operation means for adding the operation result to the bit array of the address data for context determination.

Here, a plurality of different types of calculations are performed by using a plurality of calculation means, and the calculation results are added to the bit array of the address data for context determination. Further, the value other than the pixel value of the peripheral pixel is a pixel value including peripheral pixels other than the peripheral pixel. The calculating means calculates an average value of the pixel values of the input pixels, and sets a binary or multi-value output as a calculation result. The calculating means outputs a result of the binarization based on the average density and the corresponding dither threshold as the calculation result. The calculating means outputs binary or multi-valued data of property and attribute information of the image including the edge portion and the flat portion of the image. Further, the calculating means includes: a change rate of an average density in an area of a plurality of pixels;
Calculates the magnitude, plus or minus, and outputs it. Further, the arithmetic means inputs the coordinates of the pixel of interest and converts the value into a coordinate value on a dither matrix as a calculation result, or binarized or multi-valued multi-valued data, or A table for obtaining a dither threshold from coordinate values is provided, and a dither threshold corresponding to an input or data obtained by binarizing or multi-leveling the dither threshold is output.

The image processing apparatus according to the present invention is an image processing apparatus having an image compression unit that determines a context using a peripheral pixel value of a pixel of interest as a bit array of address data and performs arithmetic coding based on the context. In the case where the context is determined by the image compression unit, the bit array of the address data for determining the context in the arithmetic coding process in the image compression, in addition to the peripheral pixel value of the target pixel, other than the pixel value of the peripheral pixel Calculate the value of
It is characterized in that the operation result is added to the bit array of the address data for context determination. Here, a plurality of different types of operations are performed, and the operation results are respectively added to the bit array of the address data for context determination.

According to another aspect of the present invention, there is provided a storage medium for storing a computer-readable program for compressing an image by arithmetic coding in an image processing apparatus, wherein the program includes at least an arithmetic coding process in image compression. As a bit array of the address data that determines the context at the time of the operation, in addition to the peripheral pixel value of the target pixel, a value other than the pixel value of the peripheral pixel is calculated, and the calculation result is converted into a bit array of the address data of the context determination. Including steps to add.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below in detail with reference to the configuration example of an image encoding unit shown in FIG.
First, data binarized from multi-value data by a dither processing unit (not shown) or the like is an input image to be compressed by the main image encoding unit.

Each pixel of the input image is input to the delay unit 4 and the FIFO 1 in raster order. The output of the delay unit 4 is
It is output after a delay of 0 or more clocks from the input. In the FIFO1, input data is delayed by one horizontal line and output as data. The output of FIFO1 is input to delay unit 5 and FIFO2. FIFO
2 is input to the delay unit 6 and the FIFO 3, and
The output of FO3 is input to delay unit 7. Delay unit 5,
The functions of 6 and 7 are the same as those of the delay unit 4, and
Functions 2 and 3 delay one line as in FIFO1.

The maximum delay amount of the delay units 4, 5, 6 can be set larger than that of the delay unit 7, and the maximum delay data of the delay unit 7 becomes the target pixel. Therefore, the outputs of the delay units 4 to 7 are pixel values around the target pixel, the delay unit 6 outputs data one line before, the delay unit 5 outputs two lines before, and the delay unit 4 outputs data three lines before. The difference between the maximum delay amount of the delay unit 7 and the delay amounts of the delay units 4, 5, and 6 corresponds to a shift of the horizontal position from the target pixel.

[0015] Of the plurality of pixels around the target pixel, m
A pixel (m bits) is input as an address of the reference memory 9. Similarly, Q1 pixel values are supplied to the arithmetic unit 8-1 by Qn.
The pixel values are input to the calculation unit 8-n. Arithmetic unit 8-1
8-n input the operation result bits P1 to Pn as address lines of the reference memory 9. Therefore, the reference memory 9 can store data in 2 ^{(m + P1 +} ... ^{+ Pn)} contexts determined by the address of (m + P1 +... + Pn) bits.

As described below, the operation units 8-1 to 8-8
−n is not limited to the one that calculates the pixel value, and is not limited to the address of the target pixel or the combination thereof, as long as the predictive accuracy of the target pixel is improved. , The encoding compression rate and the encoding speed are improved. The arithmetic code unit 10 performs an arithmetic code operation on the target pixel value and data such as the state number stored in the reference memory 9 as described below.

In the arithmetic code operation process, the pixel value of interest is dominant (sex) or inferior (sex) in appearance probability according to the state number or the like.
The number of the probability value is calculated from a table in the arithmetic coding unit 10 and a code is generated based on the data. In general, the closer the probability of the superior symbol is to 1, the smaller the amount of code can be used for encoding, and no inferior symbol is generated.

The arithmetic code unit 10 generates a code depending on whether the input pixel value of interest is a superior or inferior symbol.
The appearance probability of each of the superior and inferior symbols changes depending on the input target pixel value, or the superior and inferior symbols change places. The arithmetic coding unit 10 updates the contents of the result obtained from the table in the arithmetic coding unit 10 by writing the new state number and the dominant symbol value in the reference memory 9 and constantly improving the efficiency of the arithmetic code in a new context state. It is a mechanism to improve.

FIG. 2 shows an example of a positional relationship between a reference pixel input directly to the reference memory 9 as an address, an operation pixel input to the operation units 1 to n, and a pixel of interest. As shown in FIG. 2, the calculation pixels need to include pixels that are not included in the reference pixels. In addition, the operation units 1 to n can also increase the effect by inputting information other than the illustrated operation pixels or holding the information in advance.

(Example 1 of Operation Unit) The operation unit obtains an average value of a plurality of input pixel values and outputs the average value with a plurality of bits. Alternatively, a result of comparing the average value of a plurality of inputs with the threshold value Th1 is output in one bit. At this time, the prediction accuracy of the pixel value of interest is improved by providing the arithmetic unit with a wider range of pixels than the m-bit reference pixels provided to the reference memory 9. This is because the difference between the local average around the target pixel and a wider average level improves the accuracy of determining whether the target pixel is a flat density portion or an edge portion.

Further, if the average value is quantized using 2 to 3 bits, more accurate prediction of the target pixel can be performed. (Example 2 of operation unit) Data other than the pixel value around the target pixel is given to the input of the operation unit. In this example, the horizontal of the pixel value of interest,
Gives the vertical pixel position. Further, a dither threshold matrix is provided in the operation unit. This dither threshold matrix is matrix data used when the binary image data to be compressed is generated from the multi-valued data. By outputting the dither threshold data corresponding to the target pixel position, The prediction accuracy can be improved.

In order to reduce the number of contexts,
The threshold value is compared with another threshold value Th2, and is output as 1-bit data. Further, another threshold value Th3, Th4, Th5
This is also effective as an output of the arithmetic unit as 2-bit quaternary data as compared with. Furthermore, from horizontal and vertical positions, R × S
The remainder of the horizontal position divided by R and the remainder of the vertical position divided by S, Sv, are used as the output of the calculation unit as the positions on the dither matrix, and the number of bits is further reduced by quantizing Rh and Sv to 1 or 2 bits. It is effective even if dropped. In this case, only the size of the dither threshold matrix is required, and no threshold data is required.

Further, as a form in which the above configuration is developed,
As described above, the average value obtained in Example 1 of the arithmetic unit is binarized with the dither threshold value obtained, and the result is output to the arithmetic unit. improves. Further, when the arithmetic units described in Example 1 of the arithmetic unit and Example 2 of the present arithmetic unit are used together with the arithmetic units 8-1 to 8-n, the compression efficiency is further improved.

(Example 3 of Operation Unit) The operation unit 10 multiplies the pixel value around the target pixel once to perform primary differentiation and secondary differentiation, and obtains the result of processing such as obtaining the edge strength. Or the binary data is subjected to pattern matching with previously created data to identify whether the peripheral pixel area is an edge portion or a flat portion, and outputs the identification result as 1-bit data from the arithmetic unit. Even when the processing is performed as in this example, an improvement in prediction accuracy can be expected.

Further, the identification result is set to about 2 to 3 bits,
By discriminating and outputting the intermediate property between the edge and the flatness, the prediction accuracy as a whole is further improved. (Example 4 of Calculation Unit) This example is an example in which the magnitude of the change rate of the average density in the area of a plurality of input pixels is used as output data, and is used together with the calculation unit for calculating the above-described average density value and dither threshold. is there. Further, when the positive and negative values of the change rate of the average density are added as the calculation result, the effect is further improved.

In the above-described examples of the operation units, the input image is 2
Value image, but each pixel of the input image is quaternary,
It goes without saying that 2-bit multivalued data may be used. In this case, the following examples of the operation unit are also effective in improving the prediction of the prediction pixel. That is, it is determined whether or not the two bits forming each pixel match, and the number of matches is binarized or quantized into multi-values and output. By performing such processing, it is possible to improve the target pixel prediction by determining whether or not binary data having a background portion such as a margin of an image or a gamma characteristic of the image is at an intermediate density level. it can.

FIG. 3 shows an example of a schematic configuration of a coding apparatus using predictive coding, which will be briefly described. It is obvious that the decoding has a similar configuration. In the figure, reference numeral 10 denotes an arithmetic coding unit similar to that of FIG. 1, and the binary data PIX of the pixel of interest to be input is input to the exclusive NOR gate 104. Also, binary data CX (context) including m reference pixels in the vicinity of the target pixel and the output of the calculation unit is input to a reference memory (predicted state memory; RAM) 9.

The RAM 9 sets the predicted pixel data MPS (More P) to 0 or 1 depending on the state of the context CX input.
The exclusive NOR gate 104 is input to the exclusive NOR gate 104 as a robable symbol. The exclusive NOR gate 104 checks the coincidence / mismatch between the target pixel data PIX and the predicted pixel data MPS from the RAM 101, and inputs 1 to the arithmetic encoder 103 if they match and 0 if they do not match. Further, the RAM 9 outputs the predicted state value ST, converts the predicted state value ST into an estimated occurrence probability (the size of the dominant symbol) LSZ in the probability estimation table 105, and sends it to the arithmetic encoder 103.

As described later, the arithmetic encoder 103 has an interval size (Current Coding Interval).
Register and code register (Code register)
A C register as r) is provided, performs an arithmetic operation based on the input binary data, and outputs a determined code as compressed and encoded data. The value of the A register is 8000H because the value of the A register is smaller than that before the encoding operation, so that the accuracy of the next and subsequent operations is maintained.
The normalization process is performed as described above. In the normalization process, both the A register and the C register are bit-shifted.
At that time, the bit data shifted out from the most significant bit of the C register becomes the encoded code. When performing the normalization processing, the contents of the RAM 9 are also updated. The update data is generated by the prediction state update unit 102 based on the operation result of the arithmetic encoder 103 and sent to the RAM 9.

FIG. 4 shows an example of the configuration of the arithmetic encoder 103. Hereinafter, the arithmetic encoder 103 will be briefly described. In FIG. 4, reference numeral 301 denotes the aforementioned A register, 302 denotes the C register, 303 denotes a terminal for inputting the LSZ that is the estimated occurrence probability, 304 denotes a terminal for inputting 1-bit information to be encoded, and 305 denotes (A-LSZ ) Or a shift amount encoding circuit 306 for calculating the shift amount from the value of LSZ.
LSZ) or a subtraction / selector section for outputting LSZ, 3
07 is an addition / selector unit that outputs {C + (A−LSZ)} or C; 308 is a first shifter that shifts the output of the subtraction / selector unit 306 based on the shift amount output from the 305; Is a second shifter that shifts the output of the addition / selector unit 306 based on the shift amount output from the 305, and 310 is a terminal that outputs a code output that is considered to be shifted from the second shifter. Reference numeral 311 denotes a terminal that outputs an update instruction signal UPDATA to the predicted state update unit 102.

Each output of 305, 306 and 307 is 30
4 is switched according to the encoded 1-bit information (output from the exclusive NOR gate 104) input from the input terminal 4. For example, when the one-bit information is “1”, the shift amount based on the value of (A-LSZ) is obtained from 305, (A-LSZ) is obtained from 306, and ΔC + (A-LSZ) is obtained from 307.
Z)｝ is output. On the other hand, when the 1-bit information is “0”, the shift amount based on the value of LSZ is
LSZ is output from 06, and C is output from 307. As described above, it is generally considered that encoding of the shift amount, calculation of (A-LSZ), calculation of {C + (A-LSZ)}, shift processing, and sequential processing are performed.

FIG. 5 shows an outline of a conventionally known encoding processing flow, and FIG. 7 shows a general flowchart of the processing "ENCODE" of the JBIG encoding algorithm. Hereinafter, the conventional encoding operation will be described with reference to FIGS. Step S1300 is a reading process.
The prediction state ST and the prediction symbol MPS, which are values for calculating the appearance probability of the coded pixel from the AM 101, are read. Here, in the present embodiment, the address input at the time of the reading process is generated from the reference pixels around the encoding target pixel PIX and the output from the arithmetic unit. The probability estimation value decoding process in step S1301 converts the ST read in step S1300 into a probability estimation value LSZ that is a pixel appearance probability.

Next, an arithmetic operation is performed using PIX, MPS, and LSZ. In JBIG encoding, LSZ and MPS uniquely determined by a context must be adaptively updated in the course of encoding. In step S1302, it is determined from the result of the calculation α whether or not this update process needs to be performed. This processing is performed in steps S2100 and S210 in FIG.
2, and (AL) of the processing of S2101a and S2101b.
SZ). That is, the update process
This is executed when PIX is not MPS or when the result of (A-LSZ) is less than 0x8000.

When the update process is selected, the process of writing the calculation β and the calculation result is performed in step S1303 in FIG. Step S1303 is a process performed when an update process is necessary. In the process of writing to the RAM 101, the next predicted state NST in the context and the next MPS, NMPS, are written to the RAM 101. The address to be written to the RAM 101 is the context of the currently processed pixel used for the read processing. The calculation β corresponds to steps S2103a, S2103b, S2103 in FIG.
This corresponds to the processing of 2104a, S2104b, and S2109. The writing process is performed in steps S2105 to S2 in FIG.
108.

If the updating process is not required, the calculation β and the writing process are not performed, and the calculation γ of step S1304 is performed, and the process proceeds to the next pixel. The operation γ is an operation performed when the update process is not necessary, and is performed in step S2101a,
This corresponds to a portion where the result of (A-LSZ) in the processing of S2101b is substituted into the A register. FIG. 7 shows an example of an image processing apparatus that specifically uses JBIG image compression.

From the computer, the interface 10
01 is temporarily received by the temporary buffer 1002, the bitmap data drawn and developed by the drawing unit 1003 is drawn into the band buffer 1004, and the bitmap data of the band buffer 1004 is written by the encoding unit 1004. The data is compression-encoded, and the compression-encoded data is stored in the page memory 1006.

The coded data in the page memory 1006 is te-decoded by the decoding unit 1007, and the obtained bitmap data is printed out by the printer engine unit 1008. Here, when the data encoded by the JBEG method is decoded, the data output rate from the decoding unit 1007 is not constant and cannot be directly output from the decoding unit 1007 to the printer engine unit 1008. Therefore, the decoding unit 100
7 and the printer engine unit 1008
(First In First Out Memory) and the decoding unit 1
The bitmap data output from 007 is temporally smoothed before being output to the printer engine unit 1008.

The present invention can be applied to a system including a plurality of devices (for example, a host computer, an interface device, a reader, a printer, etc.), but can be applied to a single device (for example, a copier, a facsimile). Device). Further, an object of the present invention is to provide a storage medium storing a program code of software for realizing the functions of the above-described embodiments to a system or an apparatus, and a computer (or CPU or MPU) of the system or apparatus to store the storage medium. Needless to say, this can also be achieved by reading and executing the program code stored in the program. In this case, the program code itself read from the storage medium implements the functions of the above-described embodiment, and the storage medium storing the program code constitutes the present invention.

As a storage medium for supplying the program code, for example, a floppy disk, hard disk, optical disk, magneto-optical disk, CD-ROM, CD
-R, a magnetic tape, a nonvolatile memory card, a ROM, or the like can be used. When the computer executes the readout program code, not only the functions of the above-described embodiments are realized, but also an OS (operating system) running on the computer based on the instruction of the program code. It goes without saying that a case where some or all of the actual processing is performed and the functions of the above-described embodiments are realized by the processing is also included.

Further, after the program code read from the storage medium is written into a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, based on the instructions of the program code, It goes without saying that the CPU included in the function expansion board or the function expansion unit performs part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

[0041]

As described above, according to the present invention,
It is possible to provide a context generation method, an image compression method, and an image processing device for generating a reference state (context) for improving a compression ratio without increasing the number of contexts (the number of reference tables) so much. That is, the data string of the peripheral pixel value at the pixel position of interest is given as the address of the context reference table, and the pixel data in a wider range than the peripheral pixels can be processed by the arithmetic processing unit and compressed to a smaller number of bits. Also, it is converted into meaningful data (such as average density and corresponding dither threshold etc.) that contributes to compression efficiency,
By increasing the necessary minimum context, the prediction accuracy of the target pixel can be improved. Furthermore, the compression efficiency can be improved by increasing the appearance probability of the dominant symbol in each context so as to approach one.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a configuration example of an image encoding unit according to the present embodiment.

FIG. 2 is a diagram illustrating an example of a processing pixel position according to the present embodiment;

FIG. 3 is a schematic configuration diagram of an arithmetic encoding unit using general prediction encoding.

FIG. 4 is a block diagram of a configuration example of an arithmetic encoder.

FIG. 5 is a diagram schematically showing a processing flow of arithmetic coding.

FIG. 6 is a diagram illustrating processing of an algorithm of JBEG encoding “ENC
9 is a general flowchart of “ODE”.

FIG. 7 is a diagram illustrating a general configuration of an image processing apparatus using JBEG.

Claims (21)

[Claims] 1. A method for calculating a value other than a pixel value of a peripheral pixel in addition to a peripheral pixel value of a pixel of interest as a bit array of address data for determining a context in arithmetic code processing in image compression. A method for generating a context, comprising adding a result to a bit array of address data for context determination. 2. The context generation method according to claim 1, wherein a plurality of different types of operations are performed, and the operation results are respectively added to a bit array of address data for context determination. 3. The method according to claim 1, wherein the value other than the pixel value of the peripheral pixel is a pixel value including a peripheral pixel other than the peripheral pixel. 4. The context generating method according to claim 3, wherein in the calculation, an average value of the pixel values of the input pixels is obtained, and a binary or multi-value output is used as a calculation result. 5. The method according to claim 3, wherein a binarized result based on the average density and the corresponding dither threshold is output as the calculation result. 6. The method according to claim 3, wherein in the calculation, the property and attribute information of the image including the edge part and the flat part of the image are output as binary or multi-valued data. 7. The method according to claim 3, wherein in the calculation, a change rate, a magnitude and a sign of the average density in an area of a plurality of pixels are calculated and output. 8. In the calculation, the coordinates of the pixel of interest are input, and as a result of the calculation, values converted into coordinate values on a dither matrix, binary or multi-valued data, or coordinate values on a dither matrix are calculated. 3. The context generating method according to claim 1, further comprising a table for obtaining a dither threshold value, wherein a dither threshold value corresponding to an input or data obtained by binarizing or multi-leveling the dither threshold value is output. 9. In an image compression method for determining a context by using a peripheral pixel value of a target pixel as a bit array of address data and performing arithmetic coding based on the context, a context for arithmetic code processing in image compression is determined. In addition to the peripheral pixel value of the target pixel, a value other than the pixel value of the peripheral pixel is calculated as the bit array of the address data to be calculated, and the calculation result is added to the bit array of the address data for context determination. Image compression method. 10. The image compression method according to claim 9, wherein a plurality of different types of operations are performed, and the operation results are respectively added to a bit array of address data for context determination. 11. An image compression apparatus comprising: context determining means for determining a context by using a peripheral pixel value of a pixel of interest as a bit array of address data, and performing arithmetic coding based on the context; As a bit array of address data that determines the context in arithmetic coding processing in image compression, in addition to the peripheral pixel value of the target pixel, a value other than the pixel value of the peripheral pixel is calculated, and the calculation result is used to determine the context. An image compression apparatus, comprising an arithmetic unit for adding to a bit array of address data. 12. The image compression apparatus according to claim 11, wherein a plurality of different types of calculations are performed by a plurality of calculation means, and the calculation results are respectively added to a bit array of address data for context determination. . 13. The image compression apparatus according to claim 11, wherein the value other than the pixel value of the peripheral pixel is a pixel value including a peripheral pixel other than the peripheral pixel. 14. The image compression apparatus according to claim 13, wherein said calculation means obtains an average value of pixel values of the input pixels and outputs a binary or multi-value output as a calculation result. 15. The image compression apparatus according to claim 13, wherein said calculation means outputs, as a calculation result, a result binarized by an average density and a corresponding dither threshold. 16. The image compression apparatus according to claim 13, wherein said calculating means outputs the property and attribute information of the image including the edge portion and the flat portion of the image as binary or multi-valued data. 17. The image compression apparatus according to claim 13, wherein said calculation means calculates and outputs a change rate, a magnitude, and a sign of an average density in an area of a plurality of pixels. 18. The arithmetic means inputs the coordinates of a pixel of interest and converts the value into a coordinate value on a dither matrix as a calculation result, or binarized or multi-valued multi-valued data, or dithered data. 13. The apparatus according to claim 11, further comprising a table for obtaining a dither threshold from coordinate values on a matrix, and outputting a dither threshold corresponding to an input, or data obtained by binarizing or multi-leveling the dither threshold.
An image compression apparatus according to claim 1. 19. An image processing apparatus comprising: an image compression unit that determines a context using a peripheral pixel value of a target pixel as a bit array of address data and performs arithmetic coding based on the context. When determining, in addition to the peripheral pixel value of the pixel of interest, a value other than the pixel value of the peripheral pixel is calculated as a bit array of address data that determines the context at the time of arithmetic code processing in image compression. An image processing apparatus wherein a result is added to a bit array of address data for context determination. 20. The image processing apparatus according to claim 19, wherein a plurality of different types of operations are performed, and the operation results are respectively added to a bit array of address data for context determination. 21. A storage medium for storing a computer-readable program for compressing an image by arithmetic coding in an image processing apparatus, wherein the program determines at least a context in arithmetic code processing in image compression. As a bit array of the address data, in addition to the peripheral pixel value of the target pixel,
A storage medium including a step of calculating a value other than a pixel value of the peripheral pixel and adding a calculation result to a bit array of address data for context determination.