JPS6386084A - Multilevel image encoding system - Google Patents

Abstract

PURPOSE:To execute a processing at a high speed and simply by binarizing a picture element group in an area with one special picture element in the extracted area as the basis of a comparison. CONSTITUTION:A multilevel image memory 1 outputs the data of the picture element to a data bus 102. The data outputted to the data bus 102 are given through a selector 3 to a register 5 and a multiplexer 4. The data on data buses 105 and 106 from the register 5 and the multiplexer 4 are inputted to a comparator 6 and compared, and the output data are fetched into mean value computing elements 7 and 8. Data buses 110 and 111 from the mean value computing elements 7 and 8 come to be a data bus 113 and the converted code is formed. The data on the data bus 113 are written to the memory cell in a memory 2 for a code.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a coding method for multi-valued images, and in particular to an irreversible coding method that cannot accurately restore the original multi-valued image. It is.

Conventional Techniques The conventional method for encoding a multivalued image into binary values of 0 and 1 is as follows.

First, a region consisting of mXn (mal, n>1) multivalue pixels is extracted from the multivalue image. Next, the average value of the m×n multivalued pixels within that area is determined. Furthermore, this average value is individually compared with all m×n multi-value pixels. As a result of the comparison, if the average value is larger than the multivalued pixel being compared, 0 is assigned. On the other hand, if the average value is equal to or smaller than the multi-value pixel being compared, 1 is assigned. Through this comparison operation, an m×n bit code is obtained from m×n multivalued pixels.

At this stage, the m×n multivalued pixels are classified into a group to which O is assigned and a group to which 1 is assigned. Therefore, find the average value for each group,
Convert that value to binary.

Through the above operations, two average values expressed in m×n bits of code and binary numbers are finally obtained. These are collectively defined as the code of the m×n multi-value pixels of interest.

Problems to be Solved by the Invention In the conventional multilevel image encoding method described above, each pixel is referenced at least twice before it is finally encoded. That is, when calculating the average value of m×n multi-value pixels within the extracted area, and when comparing the calculated average value with each multi-value pixel within the area.

There is a problem that the more times each pixel is referenced, the slower the processing speed becomes. SUMMARY OF THE INVENTION An object of the present invention is to provide a multilevel image encoding method that can improve processing speed by reducing the number of times each pixel is referenced as much as possible.

Means for Solving the Problems The multilevel image encoding system of the present invention for solving the above problems is as follows.

First, a region consisting of m×n (mal, n>l) multivalue pixels is extracted from the multivalue image. Next, one specific pixel is selected from among the extracted multi-value pixels, and that multi-value pixel is used as a reference for comparison. That is, this selected multi-value pixel is compared with the remaining (mXm-1) multi-value pixels. As a result of the comparison, if the value of the compared multi-value pixel is smaller than the value of the selected multi-value pixel, 0 is assigned. On the other hand, if the value of the multi-value pixel to be compared is equal to or larger than the value of the selected multi-value pixel, 1
Assign. Through this comparison operation, an m× bit code is obtained from m×n multivalued pixels.

At this stage, the m×n multivalued pixels are classified into a group assigned with 0 and a group assigned with 1. Therefore, the average value for each group is determined and the value is converted into a binary number.

Through the above operations, two average values expressed in m×n bits of code and binary numbers are finally obtained. These are collectively defined as the code of the mXn multi-value pixels of interest.

Function As described above, the multi-value image encoding method of the present invention is characterized in that one specific pixel is selected from among the extracted m×n multi-value pixels and used as a reference for comparison. conventionally,
Considering that all pixels are referred to to obtain an average value and the average value is used as a reference for comparison, in the present invention, the number of times all pixels are referred to can be reduced by one at this stage. Therefore, the processing speed can be improved as compared to the conventional method, and the processing can be simplified.

Example Hereinafter, the present invention will be explained with reference to the drawings.

FIG. 1 is an example of a block diagram showing the configuration of a system for executing the multilevel image encoding method of the present invention. This system is completely controlled by a control section 10.

The multilevel image memory 1 is connected to the address bus 10 from the control unit 10.
Memory address output to 0 and read request signal 200
In response, the data of the pixel specified by the memory address is output to the data bus 102.

The data on the data bus 102 is transferred to the control unit 1 by the selector 3.
It makes a selection upon receiving a select signal 202 from 0 and outputs it to data bus 103 or data bus 104.

The data output to the data bus 103 is captured and stored internally by the register 5 in synchronization with a synchronization signal from the control unit 10. The data in the register 5 is always transferred to the data bus 105.
is output to. Data bus 1 output from register 5
Either one of the data on data bus 104 and the data on data bus 104 output from selector 3 is selected by multiplexer 4 in response to select signal 202 from control unit 10 and output to data bus 106.

The data on data bus 105 and the data on data bus 106 are input to comparator 6. This comparator 6 is the controller 1
The data on the data bus 105 input by the comparison request signal 204 from 0 is compared with the data on the data bus 106. If the data on data bus 106 is smaller than the data on data bus 105, comparator 6 transfers the data on data bus 106 to data bus 107.
At the same time, O is output to the data line 109.

On the other hand, if the data on the take bus 106 is equal to the data on the data bus 105, or if the data on the data bus 106
If the data on data bus 105 is larger than the data on data bus 105, comparator 6 outputs the data on data bus 106 to data bus 108 and outputs 1 on data line 109.
Output.

The data on the data bus 107 is taken into the average value calculator 7, and the data on the data bus 108 is taken into the average value calculator 8. Each average value calculator 7.8 is internally equipped with a sequential adder, a counter, a divider, and a register. data line 10
When the data on the data bus 107 is 0, the sequential adder in the average value calculator 7 takes in the data on the data bus 107 in synchronization with the synchronization signal 205 from the control unit 10 and inputs the data on the data bus 107. Adds the data already in the file and stores the result. The counter also counts up in synchronization with the synchronization signal 205 from the control section 10. However, if the data on the data line 109 is 1, the average value calculator 7
The sequential adder and counter inside do not work.

On the other hand, when the data on the data line 109 is 1, the sequential adder in the average value calculator 8 takes in the data on the data bus 108 in synchronization with the synchronization signal 205 from the control unit 10. Addition is performed to the data already in the sequential adder and the result is held. The counter also counts up in synchronization with the synchronization signal 205 from the control section 10. However, when the data on the data line 109 is O, the sequential adder and counter in the average value calculator 8 do not operate.

In order to make the average value calculators 7 and 8 complementary in this way,
An impark 11 is provided on the data line 109 connected to the average value calculator 7.

The divider in each average value calculator 7.8 receives the division request signal 206 from the control unit 10, divides the data in each successive adder by the value in the corresponding counter, and calculates the obtained value. is set in each internal register.

The value held by the register in the average value calculator 7 is output to the data bus 110. On the other hand, the value held by the register in the average value calculator 8 is output to the data bus 111.

The data on data line 109 is also taken in and set in the least significant bit of shift register 9 in synchronization with comparison request signal 204. The data in the shift register 9 immediately before this set is shifted upward one bit at a time.

At the same time, the data of the most significant bit disappears. Data in shift register 9 is constantly output to data bus 112.

Data buses 110, 111, and 112 are combined into data bus 113 to form the converted code. The data on this data bus 113 is transferred to the address bus 101 by a write request signal 201 from the control unit 10.
The code is written to the memory cell in the code memory 2 specified by the memory address output to the code memory 2.

Next, the flow of operations will be specifically explained based on the embodiment of the multilevel image encoding system of the present invention shown in FIG. The multivalued image handled in this embodiment is composed of 8-bit data per pixel. Furthermore, the extracted area is 3
It consists of ×3 pixels.

First, the control unit 10 generates addresses for extracting pixels in a 3×3 matrix shown at the top of FIG. 1 from memory 1 to data bus 102
Readout is performed pixel by pixel. For example, the reading order is 11 components, 12 components, 13 components, 21 components, etc.
, 33 components, that is, using the illustrated example, 125.
100.136.206, ..., 118. Therefore, pixels in one area are read out by nine memory reads. These nine pieces of data constitute one processing unit.

The control unit 10 outputs a select signal 202 to the selector 3 to output the data on the data bus 102 onto the data bus 103 only when the memory is read for the first time among the nine memo IJ IJ-dos. do. Therefore, 125 is output to the data bus 103 at the first memory read. The register 5 is controlled by the synchronization signal 203 from the control unit 10.
sets this value 125 internally. Data is output to the data bus 104 in the second and subsequent memo IJ-modes.

The control unit 10 outputs a select signal 202 to the multiplexer 4 to first output the data on the data bus 105 to the data bus 106 and thereafter output the data on the data bus 104.

Next, the control unit 10 outputs a comparison request signal 204 to cause the comparator 6 to compare the data on the data bus 105 and the data on the data bus 106. When data selection as described above is performed using the select signal 202, the first selection is as follows.
Since the same data 125 is compared, they are always equal. As a result of the comparison, if the data on data bus 106 is smaller than the data on data bus 105, 0 is output to data line 109. In contrast, data bus 106
Is the data on the data bus 105 equal to the data on the data bus 105?
Or the data on data bus 106 is
If the data is larger than the data on the data line 109, 1 is output to the data line 109. By applying such a comparison rule to each pixel, a 9-bit code in a 3×3 matrix shown in the middle part of FIG. 2 is obtained.

If O is output to the data line 109 as a result of the comparison, the compared data is output to the data bus 107. This data is taken into the average value calculator 7 by the synchronization signal 205 and added to the internal successive adder, and the counter counts up. That is, data 100.111.12
0.118 is added and the counter value becomes 4.

If 1 is output to the data line 109 as a result of the comparison, the compared data is output to the data bus 108. This data is taken into the average value calculator 8 by the synchronization signal 205 and added to the internal successive adder, and the counter counts up. i.e. data 125.136.20
6.215.190 is added, and the counter value becomes 5.

The divider in the average value calculator 7 receives the data 100.111.120.11 in response to the division request signal 206 from the control unit 10.
The sum of 8 is divided by the counter value 4, and the result of the division, 112, is held in the register.

Similarly, the divider in the average value calculator 8 is controlled by the controller 10.
Data 125.136 due to division request signal 206 from
．． The sum of 206.215.190 is divided by the counter value 5, and data 174, which is the result of the division, is held in the register.

Therefore, the average value 112 of pixels assigned 0 is output to the data bus 110, and the average value 174 of pixels assigned 1 is output to the data bus 111. Furthermore, 8-bit data excluding the first bit indicating which pixel is assigned 1 or O is output to the data bus 112.

The data on these three data buses 110, 111, and 112 are combined to form the code shown at the bottom of FIG. That is, from the data of the 3.times.3 matrix in which each pixel is 3 bits shown at the top of FIG. 2, the 3-byte data at the bottom is obtained. This 3-byte data is the code of the multivalued pixel in the 3×3 matrix area. The obtained 3-byte code is sent to the address bus 101 by the write request signal 201.
It is written to the memory cell in code memory 2 specified by the above memory address.

As described in detail, the multilevel image encoding method of the present invention has the following features:
Using a specific pixel in the extracted area as a comparison standard,
The pixel group within the area is binarized. Therefore, the number of times each pixel is referred to can be reduced by one compared to the conventional case in which the average value of all pixels is used as a reference for comparison. As a result, processing speed can be increased, and processing can be simplified.

Note that the image obtained by decoding is considered to have approximately the same quality as the original image if the original image is fine.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a system for implementing the multilevel image encoding method of the present invention, and FIG. 2 is a block diagram showing the configuration of a system for implementing the multilevel image encoding method of the present invention. FIG. 12 is a diagram showing an example of the flow of operations when executed for. (Main reference numbers) 1. Multilevel image memory, 2. Code memory, 3.
- Selector, 4... Multiplexer, 5... Register, 6... Comparator, 7.8... Average value calculator, 9... Shift register, 10... Control tl, 100. 101 . . . address bus, 110, 111. 112. 113...
Data bus, 200...Read request signal, 201...Write request signal, 202...Select signal, 203, 205...Synchronization signal,

Claims (1)

[Claims] A region consisting of m×n (m>1, n>1) multivalue pixels is extracted from the multivalue image, and the multivalue pixels in the region are divided into 0 and 1.
The pixel group assigned 0 and the pixel group assigned 1
For each pixel group to which the pixel is assigned, a value representing the group is determined using a predetermined method, and the m×n bit code obtained by the binarization and the value representing each group are used to calculate the value within the area. In the multi-value pixel encoding method that configures the code of multi-value pixels, binarization selects one specific pixel within the above area, and then combines the selected multi-value pixel with the remaining (m×n-1) pixels. When the value of the compared multi-value pixel is smaller than the value of the selected multi-value pixel, 0 is assigned, and in other cases, 1 is assigned. A multilevel image encoding method that uses