KR0136120B1 - Method to retrieve half tone image - Google Patents

Abstract

When the halftone image is extracted in the chemical processing system, the threshold value of the dither matrix, which is compared with the input image, is set by smoothing the histogram of the input image.

Description

Halftone Image Extraction Method

1 is a diagram illustrating a dither matrix and an input image.

2 is a schematic diagram of a typical facsimile.

3 is a flowchart illustrating a conventional halftone image extraction process.

4 is a conventional histogram and threshold relationship diagram.

5 is a histogram and a smoothed state diagram according to the present invention.

6 is a flowchart illustrating a halftone image extraction process according to an embodiment of the present invention.

7 is a flowchart embodying a histogram smoothing process in FIG.

8 is a histogram state diagram used in an embodiment of the present invention.

The present invention relates to a halftone image extraction method in an image processing system, and more particularly, to a halftone image extraction method using a histogram equalization method.

In general, image processing systems such as facsimile machines, printers, and copiers detect a halftone image using a dither matrix. The halftone includes a real halftone and a dummy halftone, and their expression methods are also different. Since the present invention relates to the counterfeit half of the dummy, the description of the former case will be omitted and only the latter expression methods will be described.

Methods of expressing the counterfeit halftone include a systematic dither method, a random dither method, an error distribution method, an average error minimum (error variance) method, an average value limit method, and a concentration pattern method. Among the above expression methods, the systematic dither method determines the threshold value to be binarized from the position information of each pixel, and is the simplest method in which the threshold position in the matrix matching the entire image does not change. This systematic dither method includes a dot lump type such as spiral, scroll, and the like, and a dot scatter type having a high intermediate frequency of dots such as Bayer and Dot.

For example, when the input image is image data of M-type N columns (M × N) quantized to 8 bits, the method of transforming the input image into a halftone having 16 gray level values is as follows. Some pixels of the input image while scanning the four-row, four-column (4 × 4) dither matrix as shown in (1a) of FIG. 1 so as not to overlap from the top left of the input image as shown in (1b). If the brightness is larger than the corresponding dither matrix threshold (Aij), it is represented by white, otherwise it is represented by black to extract the dummy halftone image. In this case, the order of the threshold values Aij, that is, the order of the rows and columns of the threshold values, depends on the dither matrix type such as Bayer, dot, spiral, and scroll. Each threshold also has a different level.

3 is a flowchart illustrating a conventional halftone image extraction process, which will be described in detail with respect to each step operation. A predetermined control unit for controlling the facsimile center or halftone image extraction stores the M rows N columns (M × N) image data of the first buffer therein (step 3a). After storing the image data in the first buffer, the threshold values (A00, A01, .4) of the four rows and four columns (4x4) of the array as shown in (1a) stored in the signal processing unit of the facsimile. ..., A32, A33 are stored in a predetermined second buffer (step 3b). Then, the image data row register value i and the column register value j are set to 0 (step 3c).

The integer remainder of the image data register value (i) divided by 4 is stored as the dither matrix row register value (I), and the integer remainder of the column register value (j) divided by 4 is stored in the dither matrix column register value ( J) (step 3d). In this way, dividing the image data row column register values (i, j) by 4 and storing the remaining values as dither matrix rows and column register values (I, J) uses 4 rows and 4 columns (4 x 4) dither matrix. The dither matrix threshold corresponding to each pixel in the first buffer is extracted to extract a halftone image having 16 levels of brightness.

The data of the first row of the image data row register value (i) and the image data column register value (j) column of the first buffer in which the image data of M rows N columns (M × N) are stored are 4 rows 4 columns (4 Compare the two data to see whether or not the threshold value of the dither matrix row register value (I) and the dither matrix column register value (J) column of the second buffer in which the threshold value of the dither matrix of the (Step 3e). As a result of this comparison, if the former data is large, the I-th row J-th column of the halftone image is set to white (step 3f), and if it is small, the j-th column of the Ith row of the halftone image is set to black (step 3g). . That is, the halftone image is expressed by comparing the brightness value of one pixel with the threshold value of the dither matrix of the row and column corresponding to the white value if the brightness value of the pixel is larger than the threshold value of the dither matrix. Then, it is checked whether or not the image data column register value j is smaller than N-1 (step 3h). As a result of the checking, when the image data column register value j is smaller than N-1, the value obtained by increasing 1 to the current image data column register value j is stored as the image data column register value j (step 3j). ). Then, the integer integer remainder of the image data row register value (i) divided by 4 is stored as the dither matrix row register value (I), and the integer value remainder of the image data row register value (i) divided by 4 is dithered. The above-described operation is repeated again from the step of storing the matrix row register value I (step 3d). Through this operation process, all the N column image data of one row are extracted as the intermediate image. In addition, when all halftone representations of the image data of N columns of one row are completed, the image data column register value is equal to N-1 (step 3f), so the image data column register value j is larger than N-1. It will not be small.

If the image data column register value j is not less than N-1, it is compared whether the image data row register value i is less than or equal to M-1 (step 3i). If the image data row register value (i) is smaller than M-1, the value obtained by increasing 1 to the current image data row register value (i) is stored as the image data row register value (i) and the image data row register value (i). Set to 0 (step 3k). As in the halftone tracing of the image data of the N columns, the integer integer remainder of the value obtained by dividing the image data row register value (i) by 4 is stored as the dither matrix row register value (i), and the image data column Image data is extracted as half-tone images from the remainder of the integer value obtained by dividing the register value j by four. Through the above-described operations, halftone expression with 16 levels of brightness for M row N column (M × N) image data is completed.

As described above, the conventional halftone representation method has various problems by using thresholds having a brightness difference at regular intervals regardless of the distribution of the input image. For example, as shown in the histogram of FIG. 4, unnecessary thresholds often occurred, and even at the point where the largest number of pixels (before and after the threshold 128), the interval between the thresholds is uniform with the point where the number of pixels is small, so that the effective threshold is applied. It was very difficult to produce visually excellent halftone images of 16 gradations because of this.

Accordingly, an object of the present invention is to provide a halftone image extraction method for generating a halftone image visually.

Another object of the present invention is to provide a method for efficiently applying thresholds in halftone image extraction.

In order to achieve the above object, the present invention extracts a better halftone image visually by setting a threshold value of a dither matrix compared to an input image through smoothing a histogram of an input image when extracting a mid-reduction image in a chemical conversion processing system. .

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, numerous specific details such as components of specific circuits are shown, which are provided to help a more general understanding of the present invention, and it is understood that the present invention may be indicated without these specific details. It will be self-evident to those of ordinary knowledge. In the following description of the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted.

2 is a configuration diagram of a general facsimile to which the present invention is applied, and is a central processing unit (hereinafter referred to as a CPU) 10 which collectively controls a system and performs various operations, and under the control of the CPU 10. An original scan unit 20 for scanning an original to generate an analog image signal, converting the generated analog image signal into digital image data, and outputting the image data output from the original scan unit 20; The image processing unit 30 stores the image data and the dither matrix and encodes and decodes the image data transceiver with another system, and inputs the image data stored in the signal processor 30 under the control of the CPU 10 to form a paper. A recording unit 40 for recording the data, and a transmission unit 50 for transmitting or receiving the image data stored in the signal processing unit 30 to another system through a public network under the control of the CPU 10. It consists of.

5 is a histogram and a smoothing state diagram according to an embodiment of the present invention. (5a) shows a histogram and (5b) shows a state where the histogram is ideally smoothed.

FIG. 6 is a flowchart illustrating a process of extracting halftone images according to an exemplary embodiment of the present invention, which is executed by the CPU 10 among the components of FIG. 2. The CPU 10 first stores image data of M rows N columns (M × N) scanned from a document or received from another system and quantized in n bits (step 6a). The histogram of the stored video data is obtained, and then smoothed (step 6b).

7 is a flowchart illustrating the histogram smoothing process.

Referring to FIG. 7, the histogram of the input image is obtained, and the hydrogen of each brightness (0 × 255) is set (Histo [each brightness]) in the memory (step 7a). Then, the number of pixels whose non-zero number of brightnesses, that is, Histo [each brightness], counts a non-zero number (step 7b). For example, if it is assumed that the half-tone image having 16 steps is extracted from the input image, it is checked whether the counted number exceeds 16 (step 7c). If the counted number of checks exceeds 16, the brightness having the smallest number of pixels is selected, and the value (x) is a pixel having a different brightness in the left and right directions of the brightness stored in the first memory area B1. The first and second brightnesses y and z having the number of pixels exceeding zero but not added to the number are obtained and stored in the second and third memory areas B2 and B3 (step 7e). Then, it is checked whether both the first and second brightnesses are generated (step 7f).

If both the first and second brightness are generated as a result of the check, it is checked whether the number of pixels of the two brightnesses is the same (step 7g). If the check results indicate that the two pixels have the same number of pixels, the distance closer to the brightness value (×) having the smallest number of pixels is stored in the fourth memory area B4 (step 7h). On the other hand, if the number of pixels having the two brightnesses is different, the one having the smaller number of pixels is stored in the fourth memory area B4 (step 7k). When neither the first nor second brightness is generated, the brightness value of any one of the first and second brightnesses y and z is stored in the fourth memory area B4 (step 7i). As described above, the side close to the brightness value x having the smallest number of pixels, the side having the smallest number of pixels, or the first or second brightness (y, z) are stored in the fourth memory area B4. After storing one of the generated brightness values, the number of stored brightnesses is added to the number of pixels of brightness stored in the first memory area B1, and then Histo [k] is updated to the added value. Assume that the variable of the lower brightness among the brightness values stored in the first memory area B1 and the brightness value stored in the fourth memory area B4 is k, and the higher brightness value of the two brightness values is assumed to be variable I. Histo [I] is updated to zero (step 7i). At this time, Histo [I] whose contents have been updated to 0 are ignored. Thereafter, the count is decreased by one (step 7j), and the operation is again executed from the operation of checking whether the counted number exceeds 16 (step 7c). On the other hand, if the counted number does not exceed 16, it is checked whether the counted brightness number is 16 (step 7m). If the result of the check is 16, 16 brightnesses generated through the repetition of the above-described process are set as the first to sixteenth threshold values of the dither matrix, stored in the second buffer, and the operation ends (step 7n). However, if the counted number is not 16, it is treated as an error (step 7n).

FIG. 8 is a histogram state diagram used in an embodiment of the present invention. In order to further explain the histogram smoothing process of FIG. 7, the histogram is calculated based on the assumption that the result of the histogram is the same. As shown, the non-zero pixel number is 19 brightness. According to the present embodiment, these brightnesses may be used as a threshold. Instead of using a threshold determined arbitrarily as in the prior art, 16 gray levels are set and used as actual data, so that a more efficient and suitable threshold is used.

For example, if the input image is transformed into a halftone having 16 levels of brightness values (midtone image extraction step values), even though only 16 threshold values are required, three pieces are exceeded when 19 brightness levels exist as described above. At this time, these three are combined into different brightnesses according to a predetermined condition to make 16. In addition, when the 16 images are made, it is possible to extract visually better halftone images by adjusting the difference (interval) between each brightness selected as the threshold according to the distribution of the input image.

Specifically, the brightness P2 having the smallest number of pixels 3 among the brightnesses is set as the reference brightness. The first and second brightnesses, which are darker or brighter than the reference brightness (left or right of the reference brightness) but not exceeding the number of pixels having different brightness, are detected when the hydrogen exceeds zero. As shown, there is no first brightness and the second brightness is P3. In this case, since only one of the first and second brightnesses is generated, the generated brightness value is selected, and a value 10 obtained by adding the number of pixels 7 of the selected brightness to the number of pixels of the reference brightness is the selected brightness and the reference brightness. Is updated to the number of pixels of low brightness P2. In other words, as P3 disappears, the number of pixels merges into P2, which was 3, so that P2 becomes 10 pixels, and the total number of non-zero brightnesses decreases from 1 to 18. However, since the number of pixels counting the number of non-zero brightnesses is 18, the halftone image extraction step value 16 is still exceeded. Therefore, the brightness P13 having the smallest number of pixels 4 among the brightnesses is set as the reference brightness. The first brightness and the brighter second brightness that are darker than the reference brightness are detected without being added to the number of pixels having different brightnesses. As shown, the first brightness is P12 and the second brightness is P14. In this case, since both of the first and second brightnesses are generated, the one having the smaller number of pixels among the two brightnesses, that is, P12 is selected. The value 9 obtained by adding the pixel number 5 of the selected brightness to the pixel number of the reference brightness is updated to the lower of the selected brightness and the reference brightness, that is, the number of pixels of P12. In other words, as P13 disappears, the number of pixels is united in P12, which is 5, so that P12 becomes 9, and the total number of non-zero brightnesses is reduced from 18 to 0 by 17. However, since the number of pixels counting the number of non-zero brightnesses is 17, the halftone image extraction step value 16 is still exceeded. Therefore, the brightness P20 having the smallest number of pixels 5 among the brightnesses is set as the reference brightness. The first brightness and the brighter second brightness that are darker than the reference brightness are detected without being added to the number of pixels having different brightnesses. As shown, the first brightness is P19 and there is no second brightness. In this case, the first brightness P19 is selected. The number 16 of the selected brightness is added to the number of pixels of the reference brightness and the value 16 is updated to the lower of the selected brightness and the reference brightness, that is, the number of pixels of P19. In other words, as P20 disappears, the number of pixels merges into P19, which was 11, so that P19 becomes 16 pixels and the total number of non-zero brightnesses decreases from 17 to 16. In this way, the remaining 16 brightnesses (P2, P4 to P12, P14 to P19) are set to the respective threshold values (A00 to A33 in FIG. 1A) of the dither matrix in the order of their occurrence.

Referring back to FIG. 6, a series of operation steps (steps 6c to 6k) performed after the histogram smoothing as described above, that is, M row and N column image data stored in a buffer may be performed. As a result of comparing one-to-one correspondences with thresholds in the unit of the square root, the halftone image corresponding to the same row and column is set to black when it is smaller than the threshold, and the halftone corresponding to the same row and column when it is larger than the corresponding threshold. The process of setting the image to white is the same as the operation process (steps 3c to 3k) of FIG. 3 described above.

In summary, an embodiment of the present invention uses a histogram smoothing technique for extracting an arbitrary halftone image, in which the number of pixels having the least number of pixels (excluding zero) in the histogram of the input image is the number of pixels other than zero. One or more pixels to find the left and right neighboring brightnesses that will aggregate the number of pixels of the left and right (up and down in terms of brightness) that are not aggregated to other brightness values at the same time. Good results can also be obtained when using the method of selecting the closest among the left and right brightnesses that are not aggregated to other brightness values and combining them to the pixel number of the brightness.

As described above, by smoothing the histogram of the input image and then applying a dither matrix, there is an advantage that a visually superior halftone image can be extracted.

Meanwhile, in the detailed description of the present invention, specific embodiments have been described, but various modifications are possible without departing from the scope of the present invention. That is, the type of dither matrix or the level of brightness of an arbitrary halftone image may be modified to some extent and implemented in the same manner. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined not only by the scope of the following claims, but also by the equivalents of the claims.

Claims (2)

A method of extracting a half-tone image of a predetermined step in an image processing system, the method comprising: storing M-row N-column image data scanned from an original or received from another system and quantized by n bits in a buffer, and an input image stored in the buffer Obtaining a histogram of the image, counting the total number of non-zero brightnesses as the number of pixels to be set as a threshold value in the histogram of the input image, and extracting a halftone image from a value obtained by counting the total number of the non-zero brightnesses as the number of pixels If the value is the same, the brightness is set to each threshold value of the dither matrix, and if the number of pixels counting the total number of non-zero brightness exceeds the halftone image extraction step value, the following subprocesses (1) to (3) Repeats the histogram smoothing process so as to be equal to the halftone image extraction step value, and (1) the brightness having the smallest number of pixels. Detects the first and second brightnesses that are darker or brighter than the reference brightness but not greater than zero as the reference brightness without being added to the number of pixels having other brightnesses, and (2) the first and second brightnesses If both are generated, if the number of pixels of the two brightnesses is the same, the side closer to the brightness value having the smallest number of pixels is selected. If not, the side having the smallest number of pixels is selected, and only the first and second brightnesses are generated. If the generated brightness value is selected, (3) the number of pixels of the selected brightness is added to the number of pixels of the reference brightness by updating the value of the selected brightness and the lower brightness pixels of the reference brightness and the low brightness. Is set to a predetermined threshold of the dither matrix and the number of high brightness pixels is deleted from the threshold setting target. The M rows and N columns image data stored in the buffer are compared with the thresholds in units of square roots of the intermediate image extraction step values in units of 1: 1, and corresponding to the same rows and columns when they are smaller than the thresholds. And setting the halftone image to black and setting the halftone image corresponding to the same row and column to white when the halftone image is larger than the corresponding threshold. A method for extracting a half-tone image of a predetermined step by obtaining a threshold value from a histogram of n-row quantized M-row N-column image data in a chemical conversion processing system, the method comprising: a first number of counting the total number of non-zero brightnesses in the histogram If the value of counting the number of non-zero brightnesses exceeds the image extraction step value, the reference brightness has the smallest number of pixels and is not added to the number of pixels having other brightnesses. A second process of detecting the first and second brightnesses that are darker or lighter but the number of pixels exceeds zero, and the brightness having the smallest number of pixels when the number of pixels of the two brightnesses is the same when both of the first and second brightnesses are generated; If the side close to the value is selected and not the same, the side having the smaller number of pixels is selected, and only one of the first and second brightnesses is generated. And a third process of selecting the generated brightness value, and updating a value obtained by adding the number of pixels of the selected brightness to the number of pixels of the reference brightness to the number of pixels of the lower brightness among the selected brightness and the reference brightness, A fourth step of setting the brightness to a predetermined threshold of the dither matrix; and after setting the threshold, the high brightness of the selected brightness and the reference brightness is deleted, and the total number of non-zero brightnesses is decreased by one. And a halftone image extracting step of performing a histogram smoothing process comprising a fifth process returning to the second process if it is not equal to the halftone image extraction step value.