JPH0879531A - Preprocessing method for dither image data compression - Google Patents

Abstract

PURPOSE: To improve the transmission efficiency of video data by forming a continuous dot pattern by dividing an original picture by using one dither pattern and collecting the dot values in one group, and then, binarizing and compression-coding the dot values. CONSTITUTION: An original picture is divided into an XxY dot pattern A and a dither pattern B having the same size (same numbers of rows and lines) as that of the pattern A is set. Then the dot values of the pattern A are collected in one group. Then the dot values in the same set are divided, namely, binarized by the element of the dither pattern B of the same size held by the group correspondingly to the element, namely, to a threshold and, when the dot values are larger than the element (threshold), the values are binarized into '1'. When the values are smaller than the element, the values are binarized into '0'. Then the original picture is compressed by changing the arranging order of binarized video data to be compressed entirely by arranging each binarized set in order.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a kind of preprocessing method for dither video data compression, and more particularly to a kind of video data used before advancing coding compression to reduce the quantity after compression. The present invention relates to a processing method for improving the transmission efficiency of video data.

[0002]

2. Description of the Related Art Image data is mainly generated by scanning an analog character such as a charge-coupled device (CCD) with respect to a character or a figure of an original to obtain a continuous analog signal and further converting the analog signal into an analog-digital converter. By converting and generating continuous binary digital data (1 is white and 0 is black). The most typical one is the use of a facsimile, and in order to complete the transmission of a huge amount of video data quickly, all the video data undergoes a data compression encoding process to reduce the amount of data after compression. Therefore, the transmission efficiency is improved.

The source of the video data may be a simple text original or an image or photograph having light and dark color layers. The former is a two-color scale where most of the data generated through scanning is clearly separated, in which case traditional facsimile data processing and compression is very well suited. Because all the traditional facsimiles are heavy sentences and non-weighted images, therefore, in the general facsimile setting style (default setting), all the two-color black and white style (Black / WhiteThres) is used.
a held mode). However, for photographs or images with color layers, they are not clearly two-tone, so if they are processed in a more general two-tone format, they will be unreadable by a facsimile. Therefore, all traditional facsimiles are dithered or error diffusive.
n Method) or peripheral concentration integrated redistribution method (a
rrrelative density assigum
ent of a djacent pixels-
This problem is addressed by providing a single halftone style, such as CAPIX). However, the image data generated by such a method is very unsuitable for the traditional compression coding process. This is because, if the image has a color in the background, the frequency frequency at which black dots and white dots alternate appears in the two-color black-and-white mode even if the image data generated by any half-tone mode is used. It is much more than what is generated.

The image data between the black point and the white point as described above can be converted into G among the commonly used run length coding.
One-digit coded MH (Mod
Taking as an example the case of processing with the "improved huffman", the following compression effects of the image data can be obtained: In the single-digit coded MH, a single horizontal scanning line is used as a single coded unit, and progressively progressively processed in the vertical direction to form a video image of one page. In the encoding table, as shown in FIG. 1, the redundancy of the white line stage represents the number of consecutive white point appearances, and the redundancy of the black line stage represents the number of consecutive black point appearances, and corresponds thereafter. The coded value is the result generated through the one-digit coded MH. 2. FIG. 2 shows one scanning example, and the original scanning data display converted into a digital signal through scanning of the optical member is 00001111 (1 is a white dot, 0 is a black dot), and the total number thereof is shown. The data amount is 8 bits. If this is compressed by the one-digit coded MH, the total data amount becomes 15 bits.

As can be seen from the above description, the total amount of data after the single-digit coded MH processing is increasing rather than decreasing, which clearly defeats the purpose of our compression. This phenomenon is a drawback that is encountered when transmitting drawing sheets or photographs by facsimile, and often after dithering one A4 size drawing sheet, it takes at least 7 minutes (transmission speed) from the start of transmission to the end of transmission. It takes about 9600 bps) or more. On the other hand, a single manuscript of the same size took only a few seconds, and the difference between them was great.

Therefore, each data form between black and white produces a data amount of 9 bits of 01 2 bits 010000111, which increases the data amount by a factor of 4.5. However, the purpose of the encoding we use is to compress the amount of data we transmit, not to expand it. Therefore, regardless of whether it is the one-digit coded MH, the two-digit coded MR or the MMR, it cannot be an effective solution.

For this reason, ordinary facsimiles have not met the recipient's ideal in halftone mode. Also, the generated semi-tones are greatly reduced in the possibility of occurring between one continuous black-and-white spot because the particles are rough and the particles are rough. For this reason, even if the compression efficiency at the time of encoding is increased, there is a problem that the graphic analysis of the receiver is significantly reduced, and this one mode of the facsimile, that is, the halftone mode has a great use value. I decided to lower it.

In addition to this, there is a kind of pattern run length method. This is a method that is often used in computers, and if there is continuous data, one pattern is searched for, this one pattern is shared by all, and it is one mark placed before the data, At the same time, the patterns are released and it is shown how many same patterns are in the data. As in the example on the front, if the material is 010101010101. ． ． ． ． Then, we can make the pattern 55H, calculate how many 55H there are, and thus improve the compression ratio significantly. There are many other methods, but they also have two drawbacks: Most methods are ineffective if the graphic has hierarchical changes. This is because figures of this kind usually have a relatively low discipline, and an ideal result cannot be obtained without considerably complicated mathematical operations. 2. Since it is completely different from the MH, MR, and MMR compression methods currently used, it is not possible to take advantage of the advantages and generality of the current single-digit encoding. Therefore, it is necessary to define a new encoding method. is there. That is, in the pattern run length method, how to define EOF, how to prevent data and control code from being mixed, whether there is omission of data of one jojo line when noise is present, and subsequent data can be code-deciphered. Whether or not it requires a lot of proofs at the same time. On the other hand, if the new method does not cause any problems in transmission and does not significantly improve the definite compression efficiency, the facsimile coding will be meaningless. On the other hand, the demand of users has become more and more intense at the present time, the analysis of reception is high, the transmission time is short, and the superfine mode (superfine mode) of 300 dpi or 400 dpi is achieved.
Fine mode) is reached, and a color facsimile is reached up to this point. Currently, telephone transmission speed is 9
600 Baud rate to 14400 Bau
d rate (V.17) has been reached, and as a result, appropriate progress has been made in the method of compressing data, and it is required to break through the situation in which the current facsimile can be used only in the character form. It is not possible to send a color layer or color statement yet.

[0009]

The main object of the present invention is to:
It can be said that this increases the compression efficiency of the video data by increasing the transmission efficiency of the video data. Another object is to use the method of the present invention before an information device such as a facsimile or a computer progresses the transmission of video data, and arranges the video data in a row to provide a kind of excellent continuity. It is to obtain a high transmission efficiency by significantly reducing the amount of subsequent compressed data as the video data that it has.

[0010]

SUMMARY OF THE INVENTION The present invention is a kind of preprocessing method for dither video data compression, and mainly divides an original image into X × Y continuous point image patterns arranged over the entire original image. [A], and a dither pattern [B] having the same size (same column and row) as one point image pattern [A] is provided, and each element of the point image pattern [A] is the dither pattern [B]. ] By the point image pattern [A] covered by the element of [1], it is assumed that each one point image value within the point image pattern is synchronized (in the same row or slightly in the same row) and collected to form a group. Each one group has one corresponding dither pattern [B] element,
Further, this element is used as a threshold, and the point spread values possessed in the same group are subjected to binarization processing (processing of two color scales) and grouped into a kind of binarization ('1' or '). 0 ') processing, as well as providing a pre-processing method for dither video data compression, which can provide video data compressed by traditional compression coding.

[0011]

In the preprocessing method for dither image data compression of the present invention, the original image is mainly divided into one dither pattern to form a continuous point image pattern, and each one point image pattern is synchronized (horizontally in a horizontal direction). , Vertical direction as rows, same point, same column, or slightly same row) are deposited and collected to form one group, and each one group has one dither pattern element corresponding to each other. Element is a threshold. Binary encoding processing (processing of two color scales) is performed on the point image values owned in the same one set, and a kind of binary encoding (0 or 1) image data is grouped. Form. The video data generated by this method has a very good continuity, and most of the video data values of the same set are 1 or 0. The amount of compressed data is reduced as compared with the amount of data that has not been processed by this method and is directly compressed, thereby increasing the transmission efficiency of video data.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS As shown in FIG. 3, in order to carry out the processing method of the present invention, first, one piece of a figure to be scanned and transmitted is divided into continuous point image patterns [A]. Each of the elements in the point image pattern [A] is a point image value obtained by scanning the figure piece, and this point image value represents the lightness or darkness at the element position of the image on the figure piece. .

Next, the present invention is further one (only one).
The dither pattern [B] in the same column and row (same size) as the above-mentioned owned point image pattern [A] is used as a reference for dividing the owned point image pattern [A]. The meaning is that each element of this one dither pattern [B] is made to correspond to synchronization within any one point image pattern [A], that is, the point image value of the same column and the same row, or only one dither pattern. [B] can be judged, and each point image pattern [A] of the same size is covered.

In this description, each one element in the owned point image pattern [A] is a point image value directly obtained by scanning on the figure piece, and indicates the so-called brightness of the image on the figure piece. . Also, since the digital data after the analog signal obtained by scanning by a general facsimile is converted through the analog-to-digital converter has a width of 8 bits, the point image values obtained by directly scanning the figure are both 2 ⁸ 256 ash levels (its value is 0-255
Of course, the value of this total gray scale depends on the size of the width of the data of the analog-to-digital conversion), and the point image value obtained by one horizontal scanning line on the drawing piece is from FIG. As shown by the irregular wavy lines in FIG.
Values between 5 as well as the values obtained with different horizontal scan lines are all different.

As shown in FIG. 4, this is an example of a dither pattern [B] provided in practice in the present invention. This dither pattern [B] is an 8 × 8 pattern, and therefore the corresponding point image values are also 8 × 8 patterns in the same row and the same column. Since the dither pattern [B] is an 8 × 8 pattern, the 64 shared composition elements also have 64 thresholds, and these 64 thresholds have a total of 256 gray scale points. It is necessary to cut the image value.
For this reason, as shown in the figure, there is a difference of 4 gray scales between each one of the thresholds, so that 64 thresholds start from 0 and are equal to 4, 8, etc. When the value increases by the difference and reaches 252 and the value does not overlap, the user can arbitrarily set the arrangement position in the dither pattern [B].

Thresholds 0, 12 in the first column in FIG.
8, 32, 160, 8, 136, 40, 168 are examples of point image values obtained by cutting the first horizontal scanning line. When this point spread value is larger than the threshold value, '1' (white point) is output through the dither processing method, and in the opposite case,
Outputs "0" (black dot). The method of binarizing this kind of slitting process, that is, "0" or "1" is the principle of the process of the traditional dither method. Therefore, the eight thresholds in the first row can be expressed. The difference between these adjacent ones is very large, and therefore very fine figures can be cut.

The eight thresholds in the first column have a very large difference between adjacent ones, so if they are directly processed by the above-mentioned dither method to be binarized video data, 01010
The result is 101. In this case, the total amount of data is rather increased after the compression through the one-digit coded MH described in the above-mentioned conventional technique, and the anti-compression effect is provided.

However, we divide the single line and the two lines in the point image pattern [A] into two 4 × 8 patterns (shown in FIG. 5). We are 0, 32, 8, 40, 16
Notice that 8 is the width of the vibration between them, as well as not too great, so is 128, 160, 136, 168. If changed to 2 × 8 or 1 × 8 (shown in FIG. 6 or FIG. 7), it was found that there was some vibration (FIG. 6) or no vibration (FIG. 7). Therefore, the present invention provides the achievement of the purpose of increasing the video data compression rate by operating this characteristic.

Taking the 1 × 8 point image pattern [A] as an example, the processing method of the present invention deposits the point image values of each one point image pattern [A] (in the same row or in the same row). Corresponding to the element of the dither pattern [B] of the same size that each group has, that is, the threshold, and further divides the possessed point image value in the same set by this element, that is, 2 A binarization (0 or 1) process is performed, and when this point spread value is larger than the element (threshold), it is binarized to be 1, and in the opposite case, it is set to 0. After that, the respective binarized sets are sequentially arranged together, and at this time, the order of the binarized video data for which the original overall compression is desired to be changed. For this reason, there is a possibility that all black points or all white points will be formed in each generated set, and the video data generated by this method has extremely good continuity. This result is most suitable for compression of single digit coded MH.

As is clear in the series of eight drawings shown in FIGS. 8 to 15, in this type of processing method, if cutting is progressed by an 8 × 8 dither pattern,
The results produced are represented in the order of these eight drawings by the top to bottom eight horizontal scanning lines on the figure. The results from the start of scanning to the completion of the processing of the present invention will be described with reference to FIG. 8 by taking the point image values in the present invention as an example of 8 rows and 8 columns, and combining the point image values of FIG. 4 described above. The processing method of the present invention will be described below.

First, the eight thresholds in the first column in the dither pattern [B], 0, 128, 32, 160,
8, 136, 40, 168 (continuous rectangular wavy lines in FIG. 8), all point image patterns [A] on the same horizontal scanning line
Binarization processing is performed on the synchronous point image values (continuous irregular wavy lines in FIG. 8) therein. When this waveform line value, that is, the point spread value is larger than the threshold value, it is set to "1" by binarization, and on the contrary, it is set to "0" by binarization. This is the first, 9, 17 ,. ． ． All the point image values of the points have the synchronization element 0 in the dither pattern [B] as the threshold value, and the second, 10, 18 ,. ． ．
The point image value of the point is the synchronization element 12 in the dither pattern [B].
It means that 8 is the threshold. By analogy with this as well, all the binarized video data with a threshold value of 0 are deposited into the first group, and all the binarized video data with a threshold value of 128 are deposited into the second group. Are sequentially processed into 8 groups. For this reason, most of each one set is video data of "1" or "0" and has very good continuity, and as a result, the data amount after the single digit coded MH compression is significantly reduced. It is possible to improve the transmission efficiency.

Further, as shown in the seven drawings of FIGS. 9 to 15, we compare the point image values (irregular wavy lines in the figures) obtained by the same horizontal scanning line as an example. And the point image values obtained with the same horizontal scanning line have different thresholds in seven columns (rectangular wave lines in the figure, and different in each figure)
Although the binarized data finally completed at the lowermost part after passing through the division by each other and belonging to the same set, the sequence order changes due to injecting different sets, but rather contiguous continuity The amount of compressed data obtained by newly processing the video data having good continuity with the processing method of the present invention and further by compressing the facsimile is higher than that of the data directly compressed without the processing of the present invention. It has decreased five times.

When the present invention is actually executed, on the transmitting side, the compressed video data to be finally sent out is already different in the arrangement order and the order at the time of scanning, and is completed in accordance with each processing order. Focusing on the group, as shown in FIGS. 16 to 21, the output of one figure piece original (FIG. 16) into four groups through the processing of the present invention, but not yet overlaid Results (Figure 17)
Then, FIGS. 18 to 21 are enlarged views of each one group, and it can be understood from the enlarged view that all the light and dark portions in the same group have considerably good continuity. Therefore, it is possible to quickly transmit each group separately. However, this is the form that we have when transmitting,
In order for the receiver to obtain the complete original drawing after receiving, one control code (cont
roll code) is added, and the receiving party is informed that the received data is compressed data divided into several groups. After that, one-digit coded MH or two-digit coded MR
The transmission of the compressed video data that occurs after the above is started. When receiving only one control code, the receiving method may reduce the received compressed data to the original arrangement order again. Since this control code is not difficult even for the current programmed hardware equipment, is a completed technology, and is beyond the scope of the present invention, detailed description thereof is omitted here.

Finally, the method of the dither pattern [B] provided in the present invention will be described. First, the size of this one dither pattern [B] is the same as that of the point image pattern [A] which is divided on the drawing piece and has a continuous relationship, and further, the size of the dither pattern [B] and the dither pattern [B].
Must be in the same column and in the same row.

Next, the number of compositional elements possessed by the dither pattern [B] (the number of 64 elements in the case of an 8 × 8 pattern) is always after being completely cut through analog-digital conversion. Total gray scale (If the analog-to-digital conversion data width is 8 bits, the total gray scale is 2 ⁸
= 256). This means that if the dither pattern [B] is an 8 × 8 pattern, its 64 compositional elements must be completely cut into 256 ash grades, and The values of the 64 composition elements are arranged in a relationship of increasing by equal difference, and the distribution of 256 ash levels is completed and ended, and if the number of matrices of this one dither pattern [B] is modified, For example, the difference between each composition element (threshold) and the total gray scale have the following mathematical relationship: (If the data width after analog-digital conversion is X bits, the total gray scale = 2 ^x ) Difference between adjacent thresholds = total gray scale / total amount of thresholds (total number of elements = number of rows x number of columns). Finally, the arrangement method of the possessed elements in the dither pattern [B] is one of the features that the user can set the arrangement by himself.

[0026]

The video data generated by the preprocessing method for dither video data compression of the present invention has extremely good continuity, and most video data values of the same set are 1 or 0. The kind of image data is subjected to the traditional compression code compression, and the amount of compressed data is reduced as compared with the amount of data directly compressed without being processed by this method, thereby improving the transmission efficiency of the image data.

[Brief description of drawings]

FIG. 1 is a partial encoding table by a one-digit encoding MH which is a conventional technique.

FIG. 2 is a principle explanatory diagram for explaining a conventional one-digit coded MH compression method.

FIG. 3 is a diagram illustrating the principle of a method of dividing one drawing piece into a point image pattern [A] according to the present invention.

FIG. 4 is a layout diagram of cut values of a dither pattern [A] according to the present invention.

FIG. 5 is a layout diagram of cut values of a dither pattern in the first embodiment of the present invention.

FIG. 6 is a layout diagram of cut values of a dither pattern according to a second embodiment of the present invention.

FIG. 7 is a layout diagram of cut values of a dither pattern according to a third embodiment of the present invention.

8 is a processing result display diagram of FIG. 7. FIG.

FIG. 9 is a view showing a result of cutting of lattice point values possessing different cutting values on one horizontal scanning line according to the present invention.

FIG. 10 is a diagram showing a result of cutting of lattice point values possessing different cutting values on one horizontal scanning line according to the present invention.

FIG. 11 is a view showing a result of cutting of lattice point values possessing different cutting values on one horizontal scanning line according to the present invention.

FIG. 12 is a view showing a result of cutting of lattice point values possessing different cutting values on one horizontal scanning line according to the present invention.

FIG. 13 is a view showing a result of cutting of lattice point values possessing different cutting values on one horizontal scanning line according to the present invention.

FIG. 14 is a view showing a result of cutting of lattice point values possessed by different cutting values on one horizontal scanning line according to the present invention.

FIG. 15 is a view showing a result of cutting of lattice point values possessing different cutting values on one horizontal scanning line according to the present invention.

FIG. 16 is a result display diagram showing an implementation result of the present invention.

FIG. 17 is a result display diagram showing the implementation result of the present invention.

FIG. 18 is a result display diagram showing an implementation result of the present invention.

FIG. 19 is a result display diagram showing an implementation result of the present invention.

FIG. 20 is a result display diagram showing an implementation result of the present invention.

FIG. 21 is a result display diagram showing the implementation result of the present invention.

Claims (1)

[Claims] 1. A kind of pre-processing method for dither video data compression, which is mainly divided into an X × Y continuous point image pattern [A] arranged over the entire original figure, and one A dither pattern [B] having the same size (same row and in the same row) as the point image pattern [A] is provided, and each element of the point image pattern [A] is covered by the element of the dither pattern [B]. Then, the point image values of the synchronization within each point image pattern (in the same row or in the same row) are deposited and collected into one group, and each one group is composed of the elements of one dither pattern [B]. Corresponding to each other, this element is used as a threshold, and the point image values possessed in the same group are subjected to binarization processing (processing of two color levels) and grouped into a kind of binarization ('1'or'0') by traditional compression coding A preprocessing method for dither video data compression that can be used for video data compression.