JPH0453354B2 - - Google Patents

Description

[Detailed Description of the Invention] (1) Description of the field to which the invention pertains The present invention converts a halftone facsimile signal obtained by scanning an original containing halftone characters, halftone dots, etc. into a halftone signal suitable for encoding. The present invention relates to a halftone facsimile signal processing method for processing and converting halftone facsimile signals.

(2) Explanation of conventional technology Most of the halftone and color paintings around us are printed materials using halftone photographs. Halftone photographs are used in the printing field to express the shading of manuscripts, and the size of the ink dots gives the impression of shading to the human eye. Currently, halftone photographs are used for most of the printed materials that contain shading, such as photographs, and the period of ink dots varies from very fine to coarse. When scanning printed matter, which is a collection of dots with such periodicity, using a facsimile machine, etc., in the case of documents such as ordinary photographs, the gradation changes are sufficiently large compared to the sample density, so the gradation changes between adjacent pixels. However, in a halftone photograph, since the image signal is a collection of minute black dots having a density approximately equal to the sample density, the obtained image signal has a sharp change in gradation almost for each pixel. For image signals that do not include halftone dots, such as text and photographs, the image signal is divided into areas with sharp gradation changes such as outlines and characters, and other areas with gentle gradation changes, and then averaging processing is performed. However, for image signals that include halftone dots, due to the nature of gradation changes in halftone dots, it is necessary to remove redundancy without sacrificing definition. It has the disadvantage that it is difficult to average the data, and that it is not compatible with existing encoding methods, which assume that the image signal after processing changes in gradation like a normal photograph.

(3) Purpose of the Invention In order to solve these drawbacks, the present invention identifies halftone dots in an image signal through simple processing, and then
The present invention relates to a halftone facsimile signal processing method for converting into a halftone signal suitable for encoding without impairing definition, and will be described in detail below with reference to the drawings.

(4) Description of structure and operation of the invention As a specific example, one embodiment will be described in detail using figures.

FIG. 1 shows an example of the position of a block for determining the average signal level difference between adjacent blocks in order to identify the contours of halftone dots and halftone portions. In the figure, P ₀ , P _A , and P _B represent the average signal level of each block. If either |P ₀ −P _A | or |P ₀ −P _B | is larger than a predetermined value, P ₀
A block having an average signal level of is determined to be a contour portion. For example, if the threshold value for contour determination is 4, |P ₀ −P _A |=|6−12|=6≧4, and in the case of Figure 1, the block with P ₀ is determined to be a contour. Ru. As a result of this determination, for a block determined to be a contour, the pixels within the block are divided into groups of pixels whose average signal level is smaller than P ₀ and pixels whose average signal level is larger than P 0 , and the average signal levels P ₁ and P ₂ of each group are calculated. The signal levels of the pixels in each group are replaced by P ₁ and P ₂ , respectively. Furthermore, among the blocks that are determined not to be contours, there are still blocks such as text portions that have severe gradation changes and whose fineness should be preserved. Therefore, it is necessary to identify blocks in which the gradation changes drastically within the block. FIG. 2 schematically shows an example of level changes within a block when an image with periodic density changes is viewed as a block with a size of four pixels square. In the figure, black parts represent pixels with a relatively high density level, and white parts represent pixels with a relatively low density level. There are various other patterns, but they are omitted here. The patterns shown in Figure 2 have large gradation changes within the block, but among these patterns, the numbers in the figure are (1, 3), (3, 1), (1, 4). ,(4,
The pattern 1) is likely to appear in characters, line drawings, etc., and its fineness should be preserved.
On the other hand, patterns with large numbers (i, j) are halftone dot patterns that have a negative effect on encoding, etc., as described in the explanation of the conventional technology, and averaging processing is performed even if there is a large gradation change unless it is an outline part. There is a need.

FIG. 3A shows an example of the order in which pixels within a block are accessed in the main scanning direction in subsequent processing, and FIG. 3B similarly shows an example of the order in which they are accessed in the sub-scanning direction. Pixels in a block with density level distribution as shown in Figure 2 are accessed in the main scanning direction and sub-scanning direction in a predetermined order as shown in Figure 3, and the level difference between two consecutive pixels is calculated. Calculate. Assuming that the level of the l-th pixel accessed is Ll and the level of the l+1-th pixel is Ll ₊₁ , the difference is Dl = Ll - Ll ₊₁ . Similarly, the level difference between the l+1st pixel and the l+2nd pixel is Dl ₊₁ =Ll ₊₁ -Ll ₊₂ . Paying attention to the successive positive/negative changes of D _K and D _K+1 , in the order of Figure 3, the numbers (i,
The larger j) is, the greater the number of changes from positive to negative and from negative to positive in both the main scanning direction and the sub-scanning direction. In the signal of the level difference obtained as a result of accessing the 16 level patterns in Fig. 2 in the main scanning direction and the sub-scanning direction in the order shown in Fig. 3, the difference between the main scanning direction and the sub-scanning direction is determined. FIG. 4 shows the number of negative changes (number of changes in the main scanning direction, number of changes in the sub-scanning direction). However, the number of changes shown in FIG. 4 does not include the number of changes from 4 to 5, 8 to 9, and 12 to 13 in the access order shown in FIG. As can be seen from the figure, in the block (1, 1), there is no positive/negative change even once in both the main scanning direction and the sub-scanning direction, that is, [0, 0].
In the block (4, 4), there are 12 changes in both the main scanning direction and the sub-scanning direction. By focusing on this number of times, the complexity of the concentration distribution in the block can be detected. For example, in the example in Figure 2 (2, 2),
(2,3), (2,4), (3,2), (3,3), (
3,
4), (4,2), (4,3), and (4,4) as being caused by halftone dots, from Figure 4, for example, the number of positive/negative changes. If the threshold value is 4 times in both the main scanning direction and the sub-scanning direction, it is possible to identify the block as having occurred 4 times or more in both the main scanning direction and the sub-scanning direction. Therefore, the previous halftone dot identification is performed on blocks that are determined not to be contours, etc., and for blocks that are identified as halftone dots, each pixel in the block is divided into several pixels so that the difference in signal level between adjacent pixels is reduced. Replace it with that signal level. In other words, it is replaced with multiple signal levels P _{0 ′} , P ₀ ″, etc. having values near the average signal level P 0. All pixels of course are replaced with the average signal level P 0
May be replaced with P ₀ . In this case, the above-mentioned several signal levels may be considered as one signal level _P0 . For blocks that are not determined to be halftone dots, find the maximum signal level Lmax and minimum signal level Lmin within the block, and calculate the difference (Lmax -
Find Lmin). If the difference (Lmax - Lmin) is larger than a predetermined value, the block is determined to be a character part, etc., and the pixels in the block are replaced with P ₁ and P ₂ in the same way as for contours. Also,
If (Lmax−Lmin) is smaller than a predetermined value, the pixels in the block are set to the average signal level.
Replace with P ₀ . For example, if the threshold value for the tone difference is set to 100 for a signal level block as shown in Fig. 5, then for this block,
When Lmax = 140 and Lmin = 30, Lmax - Lmin = 110100, and the pixels in the block are replaced with P ₁ = 40 for pixels below P ₀ = 80, and P ₂ = 120 for pixels larger than P ₀ , as shown in the figure. The conversion process is as follows.

FIG. 6 shows an embodiment of the present invention. 1 is an input terminal for a halftone facsimile signal. 2 is a signal memory which stores the signal input from input terminal 1 as 1.
Store a block or multiple blocks. 3
is an average signal level calculator which calculates the intra-block average based on the signal from the signal memory 2 and sends it to the register. Average signal level registers 41, 42
|P ₀ -P _A | and |P ₀ -P _B | are calculated by subtractors 51 and 52 from the values of and 1-block line delay device 43, respectively. Based on the value, the contour etc. judgment circuit 6
Then, it is determined whether to replace the pixels of the block with P ₁ and P ₂ . If the determination circuit 6 determines that it is not a contour or the like, the memory address controller 7 reads out the signal from the signal memory 2 and sends it to the signal registers 81 and 82. The difference between adjacent pixels is calculated by the subtracter 9 and sent to difference signal registers 101 and 102. The positive/negative change detector 11 detects a positive/negative change, and the counter selector 12 sends the result to the main and sub-scanning direction counters 131 and 132.
It operates only when a signal indicating that a negative change has been detected is input. In the halftone dot judge 14, the counter 13
The halftone dot area is determined based on the value of 1,132. In the maximum value detector 156 and minimum value detector 157,
The maximum value and minimum value within the block are respectively detected, and a subtracter 16 calculates the difference between them. In the character determination circuit 17, when it is determined that the result of the halftone dot determination device 14 is not a halftone dot, it is determined based on the value of the subtractor 1 whether or not it is a character portion.

The comparator 18 compares the magnitude relationship between the block average signal level and each pixel, and calculates the _P1 calculation circuit 191.
and P ₂ calculation circuit 192 calculate the average value of P ₁ and P ₂ . Further, the pixel information memory 20 stores information indicating which pixels are included in _P1 and which pixels are included in _P2 , and the signal level replacement circuit 21 converts the signals in the block based on the determination information. Level replacement processing is performed, and a processed signal is output to the output terminal 22.

(5) Description of Effects As described above, according to the present invention, it is possible to convert a halftone facsimile signal into a signal suitable for encoding through simple processing without impairing definition. In the description, a square area of 4×4 pixels is used, but of course other numbers of pixels and shapes may be used. Further, in the halftone dot identification portion, the identification process can be simplified by changing the order of pixel access and the number of pixels to be accessed. Note that if the change in halftone dot identification is measured using a predetermined signal level, it will be less susceptible to the effects of image signal noise and the like. The signal level for measuring this change can be set to a constant value for all blocks, or can be changed for each block. For example, the signal level for measuring this change can be the average level of each block, in which case the lth accessed pixel is lower than the average level,
When the l+1th accessed pixel is larger than the average level, or vice versa, if the lth accessed pixel is larger and l+1
Suppose that there is a change when the number is small. in this way,
When the average level of each block is used, it is possible to follow the difference in local shading of the document and identify the halftone dot area with high accuracy. In addition, in the explanation, in blocks with small signal level differences, all pixels are replaced with the average signal level P ₀ , and in blocks with large signal level differences, the average of the group of pixels smaller than the average value is replaced.
Although P ₁ is replaced by the average P ₂ of a group of large pixels, it is of course possible to replace it with other signal levels. Furthermore, the number of signal levels to be replaced may be changed.

As shown in the main text, when P ₀ is replaced with P ₁ and P ₂ , "_0" is added to the pixel replaced by P 0 and the pixel replaced by P ₁ , and "0" is added to the pixel replaced by P ₂ . If "1" is assigned, the facsimile signal is converted into gradation information of P ₀ , P ₁ , P ₂ and its pixels are P ₀ , P ₁ ,
“0” and “1” indicate which of P ₂ is represented.
It can be divided into gradation information and resolution information, and by encoding the gradation information and resolution information respectively, it is also possible to efficiently transmit facsimile signals.

In addition, in order to further improve the encoding efficiency, in the blocks replaced by P ₁ and P ₂ ,
If the number of pixels replaced by P ₁ or the number of pixels replaced by P ₂ is smaller than a predetermined value, the block is replaced with P ₀ or replaced with P ₁ and P ₂ . If P ₀ of a corresponding block is extremely large or extremely small within the possible level range of P ₀ , that block can be replaced with P ₀ .

[Brief explanation of drawings]

Figure 1 is a diagram showing the position of the block for which the difference in the average signal level within the block is calculated, Figure 2 is a diagram schematically showing an example of the density change within the block, and Figure 3 is a diagram showing the positions of the pixels within the block. FIG. 4 is a diagram showing an example of the order in which the signals are read out. FIG. 4 is a diagram showing the number of density changes when the density changes in FIG. 2 are processed in the order shown in FIG. 3. FIG. P ₁ , P ₂
FIG. 6 shows an embodiment of the present invention. 1... Input terminal, 2... Halftone facsimile signal memory, 3... Intra-block average signal level calculator, 4
1, 42...Average signal level value register, 43...Average signal level value 1 block line delay device, 51,
52... Subtractor, 6... Contour etc. determination circuit, 7... Memory address controller, 81, 82... Signal register, 9... Subtractor, 101, 102... Difference signal register, 11... Positive/negative change detector, 12... Counter Selector, 131...Main scanning direction counter, 132...
Sub-scanning direction counter, 14...halftone dot judge, 156
...Maximum value detector in block, 157...Minimum value detector in block, 16...Subtractor, 17...Character etc. judgment circuit, 18...Comparator, 191... _P1 calculation circuit, 1
92... _P2 calculation circuit, 20...pixel information memory, 2
1 represents a signal level replacement circuit, 22 represents a processed signal output terminal.

Claims (1)

[Scope of Claims] 1. In a halftone facsimile signal processing method in which a halftone facsimile signal obtained by scanning a document in which halftone dots may exist is divided into blocks each consisting of a plurality of pixels and processed, For some pairs between two predetermined spatially consecutive pixels,
The change in signal level between those two pixels is measured, and the sum of the measured amounts only between the pairs that are continuous in the main scanning direction and the sum of the measured amounts only between the pairs that are continuous in the sub-scanning direction are measured. is calculated, and compared with each predetermined value. Based on the results, it is determined whether the position of the block on the screen is a halftone dot. If the block is determined to be a halftone dot, the Each pixel in the block is replaced with a number of signal levels so that the signal level difference between adjacent pixels is small, and for blocks that are not determined to be halftone dots, processing is performed using the signal level within the block. A halftone facsimile signal processing method.