JPS63275275A - Compression method for dot picture data - Google Patents

Abstract

PURPOSE:To efficiently compress dot picture data at a high speed so as to transmit it by taking data in a prescribed reference range for a same logical value and counting them from both ends of intermediate data in the counted range. CONSTITUTION:An area unit to be compressed and coded is set to be an integer multiple of dot area and the threshold matrixes of respective picture elements are displaced in a picture element unit in accordance with a prescribed reference. The number of data in which the same theoretical value continues is counted from the leading end or the tailing end of dot data which can be obtained by displacement at every prescribed unit, and data in the prescribed reference range from both ends of intermediate data in the counted range are taken for the same theoretical value and are counted. Thus, a high regulating ratio can be obtained, and address-conversion, data-conversion and coding in a comparator can be executed with respect to hardware, whereby high speed compression can be realized.

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field of the Invention) The present invention relates to a compression method using a non-perfect restoration method for efficiently compressing halftone image data.

(Technical Background and Problems to be Solved) Compression technology for pseudo-halftone images that express halftones using binary data is used to transmit halftone images and newspaper image data by facsimile. This is because image data is enormous, and it is inefficient to transfer the data as it is or store it in memory. Facsimile data transmission uses compression technology to efficiently transmit data, and conventionally, the Mll method (one-dimensional compression method;
modified Huffman Coding) and M
R method (two-dimensional compression method; Modified-RE8)
is being used. However, MtI method and MII method
The application of the pseudo halftone image of the method is based on the premise that the image data is rearranged according to the threshold value of the dither matrix to generate a longer run, and to efficiently transmit the character image. It is not easy to select a data rearrangement method that reduces the number of small runs, and this method is inappropriate for data compression of halftone images.

There is also a predictive encoding method that considers the position of the current pixel in the dither matrix as well as the states of the preceding m reference pixels, but increasing the number of m states makes it difficult to implement into an actual device. Even if the above-mentioned MH method and MR method were applied as they were, there was a drawback that the compression rate was poor.

(Object of the invention) This invention was made under the above circumstances,
SUMMARY OF THE INVENTION An object of the present invention is to provide a method for compressing halftone image data using a non-perfect restoration method for efficiently compressing and transmitting halftone image data at high speed.

(Means for Solving the Problems) The present invention provides a network that compares an image signal with n threshold values (an integer of 3 or more) per pixel and generates digital halftone data of n bits per pixel. This invention relates to a method for compressing point image data, and the above object of the present invention is to compress and encode the area unit to be an integer multiple of the halftone area, and to set the threshold matrix for each pixel in accordance with a predetermined standard in the pixel unit. Sort the halftone dot data obtained by this sorting, and use the number obtained by counting consecutive identical logical values from the beginning or end of the halftone data for each predetermined unit, and from both ends of the intermediate data in the counted range. This is achieved by counting data within a certain reference range, regarding them as the same logical value. In addition, the area unit to be compressed and encoded is an integral multiple of the halftone dot area, and the threshold matrix for each pixel is rearranged into the pixel unit according to a predetermined standard, and then the threshold within each pixel is Sort the value matrix in a predetermined order,
The bit-wise halftone data obtained by the sorting is divided into units of bits that are an integer multiple of the n, and two group parts of the units in which the same logical value continues from the beginning and end of the bit-wise halftone data. , count the number of units, shorten the intermediate data between the two group parts according to the coat table, and consider data in a certain range from both ends of this shortened data as the same logical value. This can also be achieved by adding additional counts.

(Operation of the invention) In this invention, a threshold matrix for converting human image data into halftone data is rearranged in advance according to a predetermined standard, and logical values are
In addition to outputting "0°" or "0°°" continuously, the halftone data is divided into units in predetermined bit units, and the units in which the same logical value continues from the beginning and end are counted, and the intermediate part data is Efficient data compression of halftone dot images is achieved by treating data in a certain range from both ends as the same logical value and additionally counting them.

Furthermore, by rearranging the unitized bit-by-bit threshold matrices in a predetermined order, the intermediate halftone dot data is shortened according to the coat table, thereby further improving the degree of compression. In other words, in this invention,
Redundancy based on the spatial correlation of images and the correlation of halftone dot patterns is utilized to realize efficient compression of halftone image data using a non-perfect restoration method.

(Embodiment of the Invention) Fig. 1 shows an outline of the method of the invention according to the flow of data, in which image data represented by 8 bits is halftone-dotted by a threshold matrix (psi matrix), and the halftone The dot data DD is compressed 3 and stored in memory 4, and then the data is expanded 5 to create 4-bit halftone dot data 6.
It is said that In this case, the present invention uses an incomplete restoration method in which the halftone dot data DD and the expanded data ED are different.

Figure 2 shows human image data 1 and threshold matrix 10.
As is clear from FIG. 4 bits per pixel are compared with 4 threshold values per pixel to generate binary halftone dot data DD of 4 bits per pixel. °1°° when higher than the threshold (
When it is low, it is set to 0' (white), and each bit threshold of the threshold matrix IO is independent of the level value of the pixel, and depends only on the coordinates of the pixel.

Here, as shown in FIG. 3 as halftone data CDI, the size of the encoding block BL is determined by halftone data DDl.
In other words, the number of halftone dot units is twice that of the original number, that is, there are 2 dot units, and an encoding method is used in which each pixel has 4 bits like PL. Therefore, since the size of the encoded block BL is 2 halftone units, the image has 100 pixels/block. Then, as shown in FIG. 4, for each block, image data 1
is binarized into a threshold value matrix 10 by J: to obtain halftone dot data DD2 as shown in FIG. Next, the halftone image data (4 bits) in the block BL are rearranged in descending order of the four threshold values for each pixel in the threshold matrix 10. These data are rearranged in a fixed order determined by the threshold matrix lO.

In other words, since the threshold value for each bit of the threshold matrix lO is fixed as shown in FIG. By rearranging the
All image data are rearranged in order of increasing (or decreasing) threshold value as shown in FIG.

Then, the rearranged 400-bit halftone data DD3 is encoded every 8 bits. In this case, the rearranged 400
The bit halftone data DD3 is scanned in units of 8 bits from the beginning, and the number of units in units of 8 bits in which all 8 bits are followed by ``0°'' is determined.This group part is called the white part W, and the length of the white part W is Since the width is between °゛O゛~゛°50°', it can be encoded with 6 bits.Next, the rearranged 400-bit halftone dot data DD3 is scanned in units of 8 bits from the end. All 8 hits “1°.

Find the number of units in 8-bit units that follow. This group part is called black part B, and the length of this black part B is also 0°°~"5
0'', it can be encoded with 6 bits.

The same is true if the beginning position of part B is the number of units from the beginning of the black part.

In addition, the data in the middle part other than the white part W and the black part B of the rearranged 400-bit halftone data DD3 is called intermediate data, and this intermediate data is the result of the above rearrangement when viewed in 8-bit units. As shown in code tables O to 3 in Figure 7, there are only 25 types of patterns, and all of these can be fixedly encoded with a 5-bit encoding code. In other words, 8-bit intermediate data can be compressed into 5-bit data. In general! When is an integer, this encoding becomes one type of pattern when performed in units of ρ×n bits (gate + 1),
It can be replaced with X bits such that 2''<(n+1)'≦2x.

Furthermore, by combining the method of expressing the white part W and the black part B of the rearranged 400-bit halftone dot data DD3 as the number of units (that is, encoding the B part and the black part with 6 bits), a more efficient method can be obtained. It is possible to perform data compression. After this, in this invention, a certain range (
For example, the white part W and the black part B are expanded by regarding 16 bits of data or data of the logical value to be inverted (the number of bits is 2 bits) as the same ethical value. This situation is shown in FIG.
Even if there is, it is regarded as all <“o” and expanded to the white part W°.

In addition, the 16-bit data BM from the end of the intermediate data for the black part B is treated as 1° even if there is any °0°' in it, and is extended to the black part B'. Compressed halftone dot data M can be obtained by ignoring the part W° and the black part B°. In this case, a portion of the rearranged halftone dot data is forcibly inverted and data compressed, so even if the data is expanded, it cannot be completely restored. However, the efficiency of data compression is increasing.

In the example shown in Fig. 8, 16 bits from each end of the white part W and black part B are expanded to have the same theoretical value. The entire range up to the position where 2 bits of 0°' appear is set as 0°, and the entire range from the end of black part B to the position where 2 bits of 0°' appears is set as 0°, and the data is compressed. Also good. When compressing actual image data, the result was approximately 179.4 when the number of inversion bits was set to 2 bits, and when the number of inversion bits was set to 4 bits, in order to obtain the same logical value from both ends of the intermediate data. A compression ratio of about 1/10.3 was obtained, and when the number of inversion bits was set to 8 bits, a compression ratio of about 1710.8 was obtained.

Furthermore, in coating the intermediate part, code table 1 to
In example 3, coat °'+1001.11010°110
1+, 11100, 11101.11110, 1111
The compression ratio can also be improved by allocating data as shown in Figure 9 to 1''.In other words, 2x
-(n+1)' (2x-(n+)
1)') type of code can be assigned as a specific pattern in the middle part or a symbol in part B.

In the above embodiment, the halftone pixel data (4 bits) in the block BL are sorted by the average value of the 4-bit threshold values for each pixel in the threshold matrix 10, or
The 4-bit threshold values may be sorted in descending order of minimum value, or may be arranged in descending order of maximum value. Also, when the threshold value of the threshold matrix 10 is equal to or greater than "'128" (intermediate value), the halftone pixel data (4
Sort the 4-bit threshold values for each pixel in the threshold matrix lO in descending order of the maximum value, and when the threshold matrix value is less than or equal to 127°, the 4-bit threshold values for each pixel in the threshold matrix It is also possible to rearrange the bits in descending order of the minimum value of the threshold value.Also, in the above, the blocks are
Although the method is described as a halftone dot unit, it can be performed on any halftone dot unit, and data compression in the intermediate portion is not necessarily necessary. Furthermore, the above-mentioned "1", "°0"
The °° data may be reversed, and the logical conversion of °1°'' and "0'' with respect to the threshold position is also arbitrary.

(Effects of the Invention) As described above, according to the data compression method of the present invention, a high compression rate can be obtained, and even with regard to hardware, encoding can be performed using address conversion, data conversion, and a comparator, so high-speed compression can be achieved. I can do that.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the method of this invention according to the data flow, FIG. 2 is a diagram showing the relationship between human image data of this invention and halftone data, and FIG. 3 is an example of halftone data of this invention. Figures 4, 6 and 8 are diagrams for explaining the method of this invention, Figure 7 is a diagram showing a code table for data compression used in this invention, and Figure 9 is an extended encoding diagram. This is a code table. 1... Image data, 2... Halftone conversion, 3... Compression,
4... Memory, 5... Extension, lO... Threshold matrix. Applicant's agent Miichi Yasugata] District 1 D] D Certain 2nd sheep 3rd drawing subryokin / DD2 Certain 5th cause D3 〆Sheep 6th drawing 7th time

Claims (10)

[Claims] (1) In a method of comparing an image signal with n threshold values (an integer of 3 or more) per pixel to generate n-bit digital halftone data per pixel, the unit of area to be compressed and encoded is a halftone dot. At the same time, the threshold matrix for each pixel is rearranged in pixel units according to a predetermined standard, and data in which the same logical value continues from the beginning or end of the halftone data obtained by this rearrangement is Halftone image data characterized in that it is used as a number counted for each predetermined unit, and that data in a certain reference range from both ends of the intermediate data of the counted range are counted as the same logical value. compression method. (2) The method for compressing halftone image data according to claim 1, wherein the predetermined standard is an average value of n-bit values of each pixel. (3) The method for compressing halftone image data according to claim 1, wherein the predetermined criterion is the maximum value of n-bit values of each pixel. (4) The method for compressing halftone image data according to claim 1, wherein the predetermined criterion is the minimum value of n-bit values of each pixel. (5) The method for compressing halftone image data according to claim 1, wherein the unit unit is a bit unit that is an integral multiple of the n. (6) If the predetermined standard is such that the average value of the n-bit values of each pixel is larger than the standard value, the order is in the order of the maximum value of the n-bit values, and the average value is larger than the standard value. 2. The method for compressing halftone image data according to claim 1, wherein if the value is smaller than the minimum value of the n-bit value. (7) In a method of comparing an image signal with n threshold values (an integer of 3 or more) per pixel to generate digital halftone data of n bits per pixel, the area unit to be compressed and encoded is a halftone dot. The area is an integral multiple of the area, and the threshold matrix for each pixel is rearranged in pixel units according to a predetermined standard, and then the threshold matrix within each pixel is rearranged in a predetermined order, and the threshold matrix is obtained by the rearrangement. The number of units is obtained by converting the bit-wise halftone data into units of bits that are an integral multiple of n, and finding two group parts of the units in which the same logical value continues from the beginning and end of the bit-wise halftone data. , and shorten the intermediate data between the two group parts according to the code table, and additionally count data within a certain reference range from both ends of the shortened data, regarding them as the same logical value. A method for compressing halftone image data, characterized in that: (8) A method for compressing halftone image data according to claim 7, wherein a look-up table is used as the code table. (9) The method for compressing halftone image data according to claim 7, wherein the data in the certain range is a number of bits. (10) The method for compressing halftone image data according to claim 7, wherein the certain range of data is the number of logical values to be inverted.