JPH0468966A - Dot background character identification device and picture signal processing unit using the device - Google Patents

Abstract

PURPOSE:To realize a binarizing device with a simple device by matching a pattern of a dot part and a character pattern drawn while using the dot part as its background so as to recognize the dot part and the character part. CONSTITUTION:When a picture signal is inputted to an input terminal 1, a block segmentation section 2 segments a block comprising a prescribed picture element number and outputs the block to an average level calculation section 3 and a comparator 4. The average level calculation section 3 calculates an average level of a picture signal in a block to be inputted and the comparator 4 binarizes the picture signal in the block by using the average level as a threshold level. The binarized picture signal in the block is inputted to a pattern discrimination section 5, in which the pattern is subjected to matching with a character pattern and a dot pattern stored in advance and logical 0 is outputted in the case of the character pattern and logical 1 is outputted in the case of the dot pattern to an output terminal 6. Thus, the binarizing device is realized with a simple device.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a halftone background character identification device used in facsimiles, copying machines, etc., which identifies characters against a background of halftone dots;
The present invention relates to an image signal processing device that uses the device to recognize characters on a halftone background.

In conventional printed matter such as technical catalogues, it is often seen that letters are written against a background of halftone dots. OC characters in such manuscripts
R (Optical Character Read
er) etc., because of the halftone dots that form the background,
Characters may not be read correctly. Japanese Patent Publication No. 6325
Publication No. 0274 discloses a method for distinguishing between text, photographs, and halftone photographs, which states that these three types can be distinguished from the distribution shapes of power spectra with respect to their spatial frequencies. The above publication also states that using this identification method, character images are binarized as they are, and photographic images are also subjected to pseudo-intermediate processing using this identification method. For point photographic images, a technique has been disclosed in which smoothing processing is performed once and then pseudo intermediate processing is performed. Smoothing processing is performed on halftone photographic images.
This is to suppress the periodic structure of halftone dots and prevent moiré caused by dither processing.

Problems to be Solved by the Invention However, when using the method for discriminating text, photographs, and halftone photographs described in the above-mentioned publication, each image area is detected by determining the power spectrum of a multilevel image signal using fast Fourier transform (FFT). Therefore, the configuration of the discriminating device becomes complicated.

Furthermore, when reading and binarizing a printed document such as a catalog where the text is shaded, the apparatus shown in Figures 1 and 4 of the above publication cannot perform the binarization process on the halftone dots. Since the halftone dot processing section performs binarization processing on the character portion, the character processing section performs the binarization processing on each character section, making the device complicated.

The present invention has been made in view of the above-mentioned problems, and uses a device for identifying character areas and halftone dot areas, and a binarization method that reads and binarizes a document in which the background of the characters is shaded using the identification device and the identification result. It is an object of the present invention to provide a halftone background character identification device which can be realized with a simple device, and an image signal processing device using the device.

Means for Solving the Problems The halftone background character identification device of the present invention performs pattern matching between a pattern of a halftone dot portion and a pattern of a character portion drawn with this halftone dot portion as a background, and identifies the halftone dot portion and the character portion. This is for identifying the character part. In addition, the recognition target area, which consists of a halftone dot area and a character area drawn with the halftone dot area as a background, is scanned in the main scanning direction to count the number of black and white changes, and this counted value is converted into a binary value using a predetermined threshold. and a sub-scanning binary value obtained by scanning the recognition target area in the sub-scanning direction, counting the number of black and white changes, and converting this counted value into a binary value using a predetermined threshold. The halftone dot area and the character area are identified by the logical product. Further, an image signal processing device using the halftone background character identification device of the present invention includes any one of the above halftone background character identification devices, a low-pass filter that passes the image signal having a predetermined frequency or less, and a The image forming apparatus preferably includes a selector for selecting either the image signal or the output of the low-pass filter, and binarization means for binarizing the output of the selector.

The halftone dot area where the active halftone dots are scattered has the characteristic that black dots are dispersed, and the character area has the characteristic that the distribution of black dots is concentrated in a straight line. If these two characteristics are examined using butter/matching, it is possible to separate the two image areas.

In addition, the number of times the identification target part changes between white and black in each direction of the main scanning direction and the sub-scanning direction is counted, and the logical product of the values obtained by binarizing the number of changes in both directions with a predetermined number of times as a threshold value. If we take , the number of changes in the halftone dot area is large in both directions, and the number of changes in the text area is small in both directions, so there is a difference in the value of the logical product, and the halftone area and the text area can be distinguished.

In addition, by using one of the above methods, the halftone dot area and the character part are identified, and in the case of the halftone dot area, the overall density is made lower than the binarization threshold density by passing it through a low-pass filter. Furthermore, after suppressing the periodic structure of the halftone dots to prevent the occurrence of moiré, the halftone dots on the ground become white because they are processed by the binarization means. In addition, in the case of characters, since they are directly binarized without passing through a low-pass filter, clear binarized values can be obtained.

Embodiments Hereinafter, embodiments of the present invention will be described with reference to the drawings. 1st
The figure shows the configuration of a first embodiment of the present invention.

The first embodiment is an apparatus for identifying halftone dot areas and character areas by pattern matching. This device includes a block cutting section 2 that cuts out an image signal input from an input terminal 1 into blocks of a predetermined size, an average level calculation section 3 that calculates the average level of the image signal of the cut out blocks, and A comparator 4 that binarizes the image signal using the average level of the image signal as a threshold, a pattern discriminator 5 that judges the output pattern of the comparator 4 to determine whether it is a halftone area or a character area, and an output terminal 6. Become.

Next, the operation of this device will be explained using FIGS. 2 and 3. Figure 2 (a) shows a pattern that is identified as a character, (b)
) indicates a pattern that is determined to be a halftone dot area.

FIG. 3 is a diagram illustrating the detection situation of halftone dot areas and character areas. When an image signal is input from the input terminal 1, the block cutting section 2 cuts out a block consisting of a predetermined number of pixels as shown in FIG. 3, and outputs it to the average level calculating section 3 and the comparator 4. The average level calculation unit 3 calculates the average level of the image signal within the input block, and the comparator 4 calculates the average level of the image signal within the input block.
binarizes the image signal within the block using this average level as a threshold. The image signal in the binarized block is input to the pattern discrimination unit 5, where it is matched with pre-stored character patterns and halftone dot patterns, and rOJ is set for text and "1" for halftone dot. is output to output terminal 6.

FIG. 2 shows an example of a character pattern and a halftone dot pattern. The character pattern is linear and clustered as shown in (a), but the halftone pattern is (
As shown in b), it is discrete. Next, the procedure for removing the halftone dots on the ground will be explained with reference to FIG.

FIG. 3(a) is a diagram showing a state in which characters indicated by diagonal lines are drawn on the background of a halftone dot area, and FIG. 3(b) is a diagram showing the determination result of a halftone dot area and a character area. In FIG. 3(a), a flock 31 and a flock 32 are blocks cut out by the block cutting section 2. In FIG. Since block 3] matches pattern B in FIG. 2, it is determined to be a halftone dot area, and block 32 matches pattern 21, so it is determined to be a character area. As shown in (b), the result of this determination is represented by halftone dot #'i "1", and the character part is represented by "0".

Next, a second embodiment of the present invention will be described using FIGS. 4 and 5. The second embodiment is a halftone background character identification device that identifies halftone dot areas and character areas based on the number of black and white changes in the main scanning direction and the sub-scanning direction. In FIG. 4, input terminal 1. Block cutting part 2. Average level calculation unit 3. Comparator 4.
The output terminal 6 has the same function as in FIG. The main scanning direction change counter 5 is a counter that counts the number of 0/1 changes in the main scanning direction of the image output from the comparator 4, and the comparator 8 uses a threshold value set in advance in the threshold value generator 9. The count value of a change of 0/1 in the main scanning direction is binarized. The sub-scanning direction change counter 10 is a counter that counts the number of times the sub-scanning direction of the image output from the comparator 4 changes by 0/1.
1 binarizes the count value of O/1 change in the sub-scanning direction using a threshold value set in the threshold generator 12 in advance.

The AND circuit 13 performs a logical product of the binarized values in the main scanning direction and the sub-scanning direction, and uses this output to determine whether the area is a halftone dot area or a character area.

Next, the operation will be explained using FIG. 5. First, the statistical characteristics of halftone images and characters will be explained. A halftone image is a repetition of dots having arbitrary density levels, and the number of black/white changes is large in both the main scanning and sub-scanning directions. On the other hand, there are almost no characters in which the number of black and white changes is large in both the main scanning direction and the sub-scanning direction.

By utilizing this feature, it is possible to distinguish between halftone dot areas and character areas.

The image signal input from the input terminal 1 is processed by the cutting section 2 at the fifth
As shown in blocks 51 and 52 in the figure, blocks consisting of a predetermined number of pixels are cut out. The average level calculating section 3 calculates the average level of the image signal within this block and outputs it to the comparator 4. The comparator 4 binarizes the image signal within the block using this image signal average level as a threshold. The binarized image signal of each block is sent to the main scanning direction change counter 5.
and the sub-scanning direction change counter IOK is output. The main scanning direction change counter 7 counts the number of changes of 0/1 along a plurality of main scanning direction lines within a block, and outputs the count value after counting.

The comparator 8 binarizes the change count value in the main scanning direction using a threshold value set in advance in the threshold value generator 9 and outputs the binarized value to the AND circuit 13 . Similarly, the sub-scanning direction change color/re 10 counts the number of changes of 0/l along a plurality of sub-scanning direction lines in the block, and the comparator II counts the number of changes of 0/l along a plurality of sub-scanning direction lines in the block, and the comparator II The sub-scanning direction change count value is binarized using a threshold value and outputted to the AND circuit 13. The AND circuit 13 receives the binary values of the changes in both scanning directions and performs a logical product.

Next, a specific example of determination will be explained with reference to FIG. Figure 5 (a
) indicates that the characters indicated by diagonal lines are drawn in the halftone dot area of the background. Block 51. The block 52 is a block cut out by the block cutting section 2.

(b) shows the determination result using the output value of the AND circuit 13. Since there are dots in halftone dot areas such as block 51, the number of changes of 0/1 in both the main scanning and sub-scanning directions is large and the judgment result is "1." However, in character areas such as block 52, the density change is large. Since there is almost no difference, the number of changes in both the main scanning and sub-scanning directions is small, and the judgment result is [0].

Next, a third embodiment of the present invention will be described using FIGS. 6 to 10. This embodiment is an image signal processing device that reads characters drawn on the background of halftone dots using the halftone dot background character identification device of the first embodiment or the second embodiment. This device has an input terminal]
The first step is to identify the halftone dot area and character area of the input image signal.
Character/halftone dot discrimination unit 1 explained in the embodiment or the second embodiment
01, a low-pass filter 102 that passes only image signals with a frequency lower than a predetermined frequency of the input image signal, and an image signal input to the input terminal 1 based on the determination result of the character/halftone determination section 101 and the low-pass filter 102. , a comparator 104 that binarizes the output of the selector 103 using a threshold value set in advance in the threshold generation section 105 , and an output terminal 106 .

Next, the operation will be explained with reference to FIGS. 7 to 10.

The image signal input from the input terminal 1 is sent to the character/halftone dot discrimination section 101. Low pass filter 1o2. It is input to the selector 103. The character/halftone dot discrimination unit 101 determines whether the image signal is a character portion or a halftone dot portion, and for example, selector 103 sends a discrimination signal of rOJ in the case of a character portion and “1” in the case of a halftone dot portion. Output to.

The low-pass filter 102 outputs an image signal, which blocks image signals having a predetermined frequency or higher, to the selector 103. Selector 1
In 03, if the discrimination signal is "0", that is, a character part, the input image signal input to input terminal 1 is selected, and if the discrimination signal is "11", that is, a halftone part, the low-pass filter 1.
Select the output of 02 and see how it comes out. Comparator 104 binarizes the output of selector 103 using a threshold value set in advance in threshold generation section 105 and outputs it from output terminal 106 .

Next, this output value will be explained. Assume that the input image signal is the image shown in FIG. 3(a).

FIG. 7 shows the change in density level along the straight line of the image including the identified blur 07 in FIG. 3 (&). Since there are two halftone dots in block 3I, the density level is high in two places, and in the entire area above the straight line of block 32I-i because all are characters. FIG. 8 shows an image signal obtained by low-pass filtering the image signal of FIG. The peak value of the halftone dot area decreases, the amplitude of the periodic structure decreases, and a gentle curve is formed. FIG. 9 shows the output signal of the selector 103. That is, as the image signal of block 31, the output of the slow-pass filter 102 shown in FIG. 8 is input to the comparator 104, and as the image signal of block 32, the input image signal shown in FIG. The threshold value shown in FIG.
As shown in the figure, this value is set to be higher than the peak value of the halftone dot area.

FIG. 10 shows a state where the comparator 104 performs binarization processing using the threshold shown in FIG. 9, and shows that the halftone dot area is removed and only the character area is output.

This allows the characters drawn on the halftone dots to be clearly read.

Effects of the Invention As is clear from the above description, (1) the present invention allows characters drawn on a halftone dot background to be drawn on a halftone dot background by changing patterns between halftone dots and characters, or by changing black and white in the main scanning direction and the sub-scanning direction. By using the number of times, it is possible to distinguish between the halftone dot area and the text part, and furthermore, by using this discrimination device, the characters drawn on the halftone dot background can be separated from the halftone dots and clearly read.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a first embodiment of the present invention,
Fig. 2 is a diagram showing an example of the pattern of the text area and the halftone dot area, Fig. 3 is a diagram explaining the characters drawn on the halftone dot background and the determination results of the halftone dot area and the text area, Fig. 4 5 is a block diagram showing the configuration of the second embodiment of the present invention, FIG. 6 is a block diagram showing the configuration of the third embodiment of the present invention, FIG. 7 is a diagram showing the density distribution of the input image signal, and FIG.
9 is a diagram showing an example of the output of the low-pass filter, FIG. 9 is a diagram showing an example of the output of the separator, and FIG. 10 is a diagram showing an example of the output of the third embodiment. 1... Input terminal, 2... Block cutting section, 3...
・Average level g output, 4. 8.11. 104...
Comparator, 5... pattern discrimination section, 6. 106
...Output terminal, 7...Main scanning direction change counter, 9
．． 12. 105... Threshold generator, 10... Sub-scanning direction change counter, 13... AND circuit, 101
. . . Character/halftone discriminator, 102 . . . Low pass filter, 103 . . . Selector. Name of agent: Patent attorney Shigetaka Awano and one other person Figure (b) Figure (C1') Figure Figure 1-Flo, 731--- Glo, 732A diagram

Claims (4)

[Claims] (1) A halftone background character characterized in that the halftone dot part and the text part are identified by pattern matching between the pattern of the halftone dot part and the pattern of the character part drawn with the halftone dot part as a background. Identification device. (2) The recognition target area, which consists of a halftone dot area and a character area drawn with the halftone dot area as a background, is scanned in the main scanning direction, the number of black and white changes is counted, and this count value is set to 2 by a predetermined threshold value. The converted main scanning binary value and the sub-scanning binary value obtained by scanning the recognition target area in the sub-scanning direction, counting the number of black and white changes, and converting this counted value into a binary value using a predetermined threshold value. A halftone background character identification device, characterized in that the halftone dot portion and the character portion are identified by a logical product of the halftone dot portion and the character portion. (3) the halftone background character identification device according to claim 1, which inputs an image signal and outputs an identification signal between the halftone dot area and the character area; An image signal processing device comprising: a selector that selects either the image signal or the output of the low-pass filter based on an identification signal; and binarization means that binarizes the output of the selector. (4) the halftone background character identification device according to claim 2, which inputs an image signal and outputs an identification signal between the halftone dot area and the character area; An image signal processing device comprising: a selector that selects either the image signal or the output of the low-pass filter based on an identification signal; and binarization means that binarizes the output of the selector.