KR20050081532A - Image binary-coding method based on local characteristics - Google Patents

Abstract

The present invention relates to an image binarization method considering characteristics of each local region included in an input image, and the binarization method according to the present invention adaptively localizes in consideration of characteristics of a pixel peripheral region to be currently processed in an input image. When a threshold value is calculated and a binary pixel value corresponding to a pixel to be processed is output by comparing the calculated local threshold value with a value of a pixel to be processed at present, when a high frequency and a low frequency component are mixed in the input image. In addition, it is possible to adaptively perform the binarization so that the low frequency component of the background region is faithfully displayed while maintaining the high frequency component of the boundary portion.

Description

Image binary-coding method based on local characteristics

The present invention relates to a binarization method for outputting a continuous-tone image as a binary value, and more particularly, to a binarization method in consideration of local region characteristics of an input image.

The continuous grayscale image is an image having each pixel value in a two-dimensional space. Binarization output devices are devices that limit output values to binary values. For example, a binarization output device based on a halftone algorithm may be used for facsimile machines, printers, digital copiers and LCD panels.

When the gradation image is input, the binarization output device simplifies and outputs the gradation value to two levels, that is, 0 or 1. To this end, the conventional binarization apparatus uses a method of masking an input pixel with a threshold value according to the position of a pixel currently being processed.

1 is a block diagram showing a conventional binarization apparatus.

Referring to FIG. 1, a conventional binarization apparatus includes a counter 100, a memory address generator 110, a memory 120, and a comparator 130.

The counter 100 outputs position information in the 2D image of the pixel I (x, y) currently input. The counter 100 outputs the position information in consideration of the fact that the continuous grayscale image is input by one pixel while moving from the upper left to the lower right of the 2D image. Thus, the counter unit 100 includes an x-axis position counter 101 and a y-axis position counter 102. The x-axis position counter 101 outputs a value corresponding to the x-axis position in the two-dimensional image of the pixel I (x, y) currently input. The y-axis position counter 102 outputs a value corresponding to the y-axis position in the two-dimensional image of the pixel I (x, y) currently input.

The memory address generator 110 generates one-dimensional memory addresses and control signals corresponding to the x-axis position information and the y-axis position information output from the counter 100. The control signal is a signal for controlling the read mode for the memory 120.

The memory 120 stores a mask threshold value corresponding to each pixel of the 2D image. When the memory address and the control signal are transmitted from the memory address generator 110, the memory 120 outputs a predetermined mask threshold value M (x, y) for the pixel I (x, y) currently input.

The comparator 130 compares the magnitude between the input pixel I (x, y) and the mask threshold value M (x, y) to compare the binary pixel value B for the input pixel I (x, y). Output (x, y) For example, when the input pixel I (x, y) is larger than the mask threshold value M (x, y), the comparator 130 outputs "1" as a binary pixel value. On the other hand, if the input pixel I (x, y) is not greater than the mask threshold value M (x, y), the comparator 130 outputs "0" as a binary pixel value.

Therefore, the image quality of the binary image output from the binarization apparatus is determined according to the resolution of the mask threshold value, the distribution of the mask threshold value, and the size of the mask stored in the memory 120. In other words, if the components of the mask threshold stored in the memory 120 are regularly arranged, a regular pattern will be formed in the image output from the binarization apparatus.

2 is a diagram illustrating a masking scheme for one channel of a black and white image or a color image.

The continuous grayscale image is an image having respective pixel values in a two-dimensional space. When the output value of the output device is limited to binary image, the continuous gray value should be simplified to two levels, that is, 0 or 1, in order to output the binary image. Therefore, the binarization algorithm is needed, and in the mask method, the threshold value is changed according to the position of the pixel currently being processed, rather than using a fixed threshold like a simple binarization. have. And the threshold values corresponding to each pixel are obtained from the threshold array.

The image quality of the finally output binary image is determined according to the resolution of the threshold constituting the mask, the distribution of the threshold, and the total size of the mask.

That is, when the components of the threshold value are arranged regularly, the output image will also be influenced by the threshold value to form a regular pattern. On the other hand, the order of the thresholds does not have regularity, and a stochastic mask may be used in which a high frequency component is included and the size of the mask is larger than that of the conventional mask. The stochastic mask includes blue noise, that is, a high frequency component, and because of the large size of the mask, many thresholds can be expressed, which has the advantage of improving gray scale expression.

3 is a diagram illustrating an output image using a stokestick mask.

FIG. 4 is a diagram illustrating a mask of a Bayer Dither having an 8 × 8 pixel size that is widely used.

It is possible to reduce the influence of regular patterns due to the use of repetitive masks only when the mask has a size larger than 64 pixels in length and width. That is, as the size of the mask increases, the binarized image having the reduced pattern may be output. However, the larger the mask size, the larger the memory capacity to store the mask threshold. In addition, since the mask includes blue noise in a form in which the threshold arrangement of the mask is diffused, the high frequency component of the input video signal is reduced by the mask threshold. Therefore, as illustrated in FIG. 4, the image quality of the boundary component, which is a visually important component of the image, is degraded. Blue noise is a noise generated when the thresholds having similar values are expressed in a distant form. Thus, a mask in which the thresholds are dispersed has a good overall image quality, but due to the diffuse threshold The same part as the boundary area has a disadvantage that cannot be expressed well.

As such, when using a conventional masking method, even when using a mask of a predetermined size or more to reduce the effect of the pattern, blue noise reduces the boundary component, which is a visually important part of the image, to reduce the effect of the text. There is a problem that the tone of the boundary region of the minute portion is damaged.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problem, and provides a binarization processing method and apparatus capable of effectively representing a boundary portion and a background region of an input image by performing a binarization process in consideration of local region characteristics of an input image. The purpose is.

In order to achieve the above objects, a binarization method for outputting an input image as a binarized image includes: calculating a local threshold value in consideration of characteristics of a pixel peripheral region to be processed currently; And comparing the detected local threshold value with the value of the pixel to be currently processed to output a binary pixel value corresponding to the pixel to be currently processed.

In addition, in the binarization method according to the present invention for achieving the above object, the step of calculating a local threshold value comprises the steps of: calculating a local mean value representing a pixel average value of peripheral pixel values of a pixel to be currently processed; Calculating the flatness of the peripheral pixel based on the peripheral pixel values of the pixel to be currently processed; Modifying the calculated local mean value based on the calculated local mean value and flatness; Preferably, the method comprises calculating a local threshold using the modified local mean value.

In addition, in the binarization method according to the present invention for achieving the above object, the step of calculating a local threshold value comprises the steps of: calculating the edge intensity of the peripheral pixel based on the peripheral pixel values of the pixel to be currently processed; Generating a mask threshold value corresponding to the pixel currently to be processed; And modifying the generated mask threshold based on the calculated edge intensity, wherein the local threshold calculating step preferably uses the modified mask threshold.

In addition, in the binarization method according to the present invention for achieving the above object, the step of calculating a local threshold value comprises the steps of: calculating the edge intensity of the peripheral pixel based on the peripheral pixel values of the pixel to be currently processed; Generating a mask threshold value corresponding to the pixel currently to be processed; Modifying the generated mask threshold based on the calculated edge intensity and the calculated flatness, wherein the local threshold calculation step preferably uses a modified mask threshold.

In addition, in the binarization method according to the present invention for achieving the above object, the step of calculating the local threshold, calculating a weighting for the local mean value based on the local mean value and flatness, and at least one of flatness and edge strength Computing a weight for the mask threshold using a, the local threshold is calculated using the following equation,

Here, Th _Loc (x, y) is an adaptive local threshold value, Avg _Loc (x, y) is a local mean value, F _Loc is a weight for the local mean value, and M (x, y) is a mask threshold value , F _Mask is a weight for the mask threshold value, Offset is an offset of the input image, F _offset is a weight for gamma correction of the offset.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

5 is a block diagram of a binarization apparatus according to a preferred embodiment of the present invention. Referring to FIG. 5, the apparatus according to the present invention includes a local windowing unit 510, a local average value calculator 520, a flatness calculator 530, and edge strength calculations. The unit 540, the mask threshold generator 550, the offset generator 560, the gamma error corrector 570, the first weight calculator 580, the second weight calculator 590, and the first multiplication The unit 600 includes a second multiplier 610, a first adder 620, a second adder 630, and a comparator 640.

The local windowing unit 510 provides pixel I (x, y) currently processed in an image of a two-dimensional space and pixel values located around the current pixel I (x, y). As shown in FIG. 6 (a), a pixel located near the current pixel I (x, y) is located at an upper left (UL, upper left) position based on the current pixel I (x, y). Pixel located at U, Upper, Pixel located at UR, Upper Right, Pixel located at L, Left, Pixel located at Upper Left Left, UUL, Upper Upper Left, pixels located at Upper Upper, Upper Upper Right, Upper Right Right, and Left Left It includes a pixel.

The pixel to the right of the current pixel I (x, y) cannot be stored by the causal system and thus is not included in the surrounding pixels. The value of the upper line of the current pixel I (x, y) among the pixels located in the periphery is a value stored in the line memory (not shown) when processing the previous line. Thus, it is provided from the line memory not shown.

The local windowing section 510 is composed of a shift register to shift the pixel value to the left whenever the binarization process for one pixel is completed. That is, UR moves to U, U moves to UL, UL moves to ULL, and the ULL value is removed. I value moves to L value. The UR value is provided from the not shown line memory. I value is a pixel value currently input.

The value I of the current pixel and the surrounding pixels UL, U, UR, L, LL, ULL, UUL, UU, UUR, and URR that are windowed to the local windowing unit 510 are local mean calculator 520. ), The flatness calculator 530, the edge strength calculator 540, and the comparator 640, respectively.

FIG. 6 is a diagram illustrating an embodiment of the local windowing unit 510 used in the binarization apparatus according to the embodiment of the present invention shown in FIG. 5.

The local average detector 520 uses the current value I of the current pixel and the values of the neighboring pixels UL, U, UR, L, LL, ULL, UUL, UU, UUR, and URR to calculate the local average value Avg _Loc. (x, y)) is detected as in Equation 1.

In Equation 1, a _ij is a directional filter coefficient and is a weight of each window pixel. This directional filter coefficient a _ij has a value of 1, 2, 1.414, 2.236, or the like depending on the distance from the current pixel. I (x, y) is the pixel value at each position in the filter.

In Equation 1, I (x, y) is a pixel value of a windowed position (UL, U, UR, L, LL, ULL, UUL, UU, UUR, URR). x-iΔx is the position of the x-axis pixel, and y-jΔy is the position of the y-axis pixel.

The local average detector 520 multiplies the pixel value (UL, U, UR, L, LL, ULL, UUL, UU, UUR, URR) by the weight a _ij corresponding to each pixel value transmitted from the local windowing unit 510. The sum is divided by the sum of the weights to calculate the local mean value Avg _Loc (x, y) of the currently windowed image. The calculated local average value is transmitted to the first weight calculator 580 and the first multiplier 600, respectively.

FIG. 7 is a diagram for describing a local mean value detector 520 used in a binarization apparatus according to an embodiment of the present invention illustrated in FIG. 5. Each pixel value transmitted from the local windowing unit 510 is input to the register unit 720, and the logic operation unit 740 performs a logical operation according to Equation 1 on the pixel values received from the register unit 720. Calculate and output the local mean value Avg _Loc (x, y).

The flatness calculator 530 rearranges pixel values of the peripheral pixels UL, U, UR, L, LL, ULL, UUL, UU, UUR, and URR transmitted from the local windowing unit 510 in size order. Then, the flatness _Loc (x, y) is calculated by averaging the difference between the adjacent pixel values. However, it is also possible to calculate the flatness of the surrounding pixels by using an existing algorithm for calculating the flatness. The calculated flatness flat _loc (x, y) is transmitted to the first weight calculator 580 and the second weight calculator 590, respectively. The flatness flat _loc (x, y) input to the second weight calculator 590 may be adjusted by the selector 532. For example, in order to maintain the brightness of the input image and the brightness of the output image, the switch of the selection unit 532 remains open, and in order to increase the contrast between the input image and the output image, the selection unit The switch of 532 is kept in a closed state.

7 is a view for explaining the flatness calculator 530 used in the binarization apparatus according to an embodiment of the present invention shown in FIG. Each pixel value transmitted from the local windowing unit 510 is input to the register unit 720, and the logical operation unit 740 rearranges the pixel values received from the register unit 720 in the order of size, and then each adjacent pixel value. Calculate and output the flatness Flat _Loc (x, y) by averaging the difference between them.

The edge intensity calculator 540 calculates a maximum value and a minimum value among pixel values of the peripheral pixels UL, U, UR, L, LL, ULL, UUL, UU, UUR, and URR transmitted from the local windowing unit 510. The difference between these maximum and minimum values is determined by the edge strength Edge _Loc (x, y). However, it is also possible to calculate edge intensities of surrounding pixels using an algorithm for calculating flatness. The calculated edge strength Edge _Loc (x, y) is transmitted to the second weight calculator 590. Edge Intensity Edge _Loc (x, y) is related to the probability that the current pixel is in the boundary area. When the current pixel is located in the boundary area of the picture or text, the edge intensity value is increased.

7 is a view for explaining the edge strength calculation unit 530 used in the binarization apparatus according to an embodiment of the present invention shown in FIG. Each pixel value transmitted from the local windowing unit 510 is input to the register unit 720, and the logic operation unit 740 detects a maximum value and a minimum value among the peripheral pixel values received from the register unit 720. The difference between the maximum value and the minimum value is determined by the edge strength Edge _Loc (x, y) and output.

The mask threshold generator 550 generates a mask threshold M (x, y) of the pixel to be currently processed. FIG. 8 is a diagram illustrating an embodiment of the mask threshold generator 550 shown in FIG. 5.

Referring to FIG. 8, the mask threshold generator 550 includes a counter 820, a memory address generator 840, and a memory 860. The counters 822 and 824 provided in the counter 820 are constituted by a 7-bit counter capable of counting positions from 0 to 63. Therefore, the size of the mask is 64 x 64 pixels. The counters 822 and 824 increase the count value when a processing completion signal or a line processing completion signal of one pixel is input from the outside. The position information output from the counter is input to a memory address generator 840, and the memory address generator 840 has one dimension corresponding to the x-axis position information and the y-axis position information input from the counter 820. Generates memory addresses and control signals. The control signal is a signal that controls the read mode for the memory 860. The memory 860 stores a mask threshold value corresponding to each pixel of the 2D image. When a memory address and a control signal are input from the memory address generator 840, the memory 860 outputs a predetermined mask threshold value M (x, y) for the pixel I (x, y) currently input.

The offset generator 560 generates an offset of the corresponding binarization device. It consists of one register for this purpose. An offset value to be generated in the register is stored, and an offset signal stored in each pixel is generated and output to the first adder 620 whenever a pixel processing completion signal is input. Typically, the offset value is set at the intermediate level of the maximum input level, for example, if the input pixel value has an 8 bit value, i.e., a range of 0-255.

The gamma error correcting unit 570 receives the brightness I (x, y) of the pixel to be processed and generates a value for correcting an input level and a human visual difference existing in the input device, and generates a value for the first adder 620. Output FIG. 9 is a diagram illustrating a gamma error correcting unit 570 according to an embodiment of the present invention shown in FIG. 5. Referring to FIG. 9, the address generator 920 receives a pixel I (x, y) currently being processed, generates an address of a gamma table corresponding to luminance of an input pixel, and generates a gamma table storage unit. 940, the gamma table storage unit 940 outputs a gamma error correction value corresponding to the input address to the logic operation unit 960, and the logic operation unit 960 outputs the data from the gamma table storage unit 940. A gamma error correction value F _offset is generated based on the gamma error value and output to the second adder.

The first weight calculator 580 calculates the local mean value Avg _Loc (x, y) calculated by the local mean calculator 520, and the flatness flat _loc (x, y) calculated by the flatness calculator 530. Calculate the weight F _Loc based on Equation 2.

Here, Function ₁ and Function ₂ are functions for _obtaining the F _Loc value, and determine the weighted F _Loc applied to the local average value based on the local average value and the flatness around the current pixel. In the present embodiment, Function ₁ is determined such that the local mean values Avg _Loc (x, y) and F _Loc are nonlinearly proportional, and Function ₂ is determined so that the flatness Flat _Loc (x, y) and F _Loc are inversely proportional. .

FIG. 10 is a diagram for describing a first weight calculator 580 used in the binarization apparatus of FIG. 5 according to an embodiment of the present invention. Referring to FIG. 10, the local mean value Avg _Loc (x, y) calculated by the local mean value calculator 520 and the flatness Flat _Loc (x, y) calculated by the flatness calculator 530 may be represented by a register unit ( 1020, and the logic operation unit 1060 is based on the local average values Avg _Loc (x, y) and flatness Flat _Loc (x, y) received from the register unit 720 and the parameters stored in the parameter table 1040. To calculate and output the weight F _Loc .

The second weight calculator 590 calculates the flatness Flat _Loc (x, y) value calculated by the flatness calculator 530 and the flatness Edge _Loc (x, y) calculated by the edge strength calculator 540. Using the equation, the mask weight F _Mask is calculated based on the equation (3).

Here, Functions ₃ and ₄ are functions for _obtaining the F _Mask value, and determine the weighted F _Mask applied to the mask threshold based on the edge intensity and flatness around the current pixel. In the present embodiment, Function ₃ is determined such that the edge intensity Edge _Loc (x, y) and the F _Mask value are nonlinearly proportional, and Function ₄ is determined so that the flatness Flat _Loc (x, y) and the F _Mask value are inversely proportional. .

FIG. 10 is a diagram for describing a second weight calculator 590 used in the binarization apparatus of FIG. 5 according to an embodiment of the present invention. Referring to FIG. 10, the flatness flat _loc (x, y) calculated by the flatness calculator 530 and the edge strength edge _loc (x, y) calculated by the edge strength calculator 540 are the register units ( 1020, the logic operation unit 1060 is based on the flatness Flat _Loc (x, y) and the edge strength Edge _Loc (x, y) received from the register unit 720 and the parameters stored in the parameter table 1040. To calculate and output the mask weight F _Mask .

FIG. 11 is a diagram illustrating a parameter region used for functions for _obtaining F _Loc and F _Mask according to an embodiment of the present invention. In FIG. 11, the values calculated by the local average calculator 520, the flatness calculator 530, and the edge strength calculator 540 become a part of the three-dimensional region. For example, in FIG. 11, A denotes a flat region having a weak border of the background, B denotes a region where the edge of the dark region is strong, and C denotes a slightly non-flat region without the edge of the dark background region.

The first multiplier 600 multiplies the local average value Avg _Loc (x, y) calculated by the local mean value calculator 520 and the weight F _Loc calculated by the first weight calculator 580, and thus, Avg. _Loc (x, y) * F _Loc is output to the second adder 630.

The second multiplier 610 multiplies the mask threshold value M (x, y) input from the mask threshold value generator 550 by the weight F _Mask calculated by the second weight calculator 590, and as a result, The M (x, y) * F _Mask is output to the second adder 630.

FIG. 12 is a diagram illustrating a multiplier used in the first multiplier 600 and the second multiplier 610 of FIG. 5, showing a binarization apparatus according to an embodiment of the present invention.

The first adder 620 adds the gamma error correction value F _Offset calculated by the gamma error correction unit 570 to the offset value input from the offset generator 560, and thus, as a result, Offset + F _offset is added to the second value. Output to the adder 630.

FIG. 13 is a diagram showing an adder used in the first adder 620 of FIG. 5 showing the binarization according to an embodiment of the present invention.

The second adder 630 is based on the input value from the first multiplier 600, the second multiplier 610, and the first adder 620 according to the equation (4) below the local adaptation threshold Th Compute _LOC (x, y).

Here, the first term represents the influence of the local mean value on the local adaptation threshold Th _Loc (x, y), and as a result, the result of performing high pass filtering on the input image. Here, if F _Loc is fixed at 1.0, the binary result of black or white is determined according to the magnitude of the average value of the pixel currently processed and the average filter, and many high frequency components are represented. If F _Loc becomes 0.7, according to Equation 4, since the local adaptation threshold has a magnitude of 0.7 times the local mean at the offset, many high frequency components disappear. Therefore, it can be seen that the size of the high pass filter can be adjusted according to the size of F _Loc .

The second term also serves to add a mask component to the local adaptation threshold Th _Loc (x, y). The value of F _Mask is a reflection ratio of a mask component with respect to a threshold value based on an offset. If the value of the F _Mask is increased, the local adaptation threshold is increased in expressing the continuous gray level due to the inclusion of a mask component.

In addition, the third term represents the overall offset value for the local adaptation threshold Th _Loc (x, y).

FIG. 14 is a diagram showing an adder used in the second adder 630 of FIG. 5 showing a binarization device according to one embodiment of the present invention.

The comparison unit 640 compares the magnitude between the input pixel I (x, y) and the local adaptation threshold Th _Loc (x, y) and compares the binary pixel value with respect to the input pixel I (x, y). value) Outputs B (x, y). For example, if the input pixel I (x, y) is larger than the local adaptation threshold Th _Loc (x, y), the comparator 640 outputs "1" as a binary pixel value. On the other hand, if the input pixel I (x, y) is not larger than the local adaptation threshold Th _Loc (x, y), the comparator 640 outputs "0" as a binary pixel value.

According to the comparison result, the value output from the comparator 640 may be changed according to the design of the system. The value output from the comparator 640 is a binary pixel value B (x, y) of the pixel I (x, y) currently to be processed.

FIG. 15 is a diagram illustrating an apparatus used in the comparator 640 of FIG. 5 illustrating a binarization apparatus according to an embodiment of the present invention.

The present invention is not limited to the above-described embodiment, and of course, modifications may be made by those skilled in the art within the spirit of the present invention.

The invention can also be embodied as computer readable code on a computer readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, hard disk, floppy disk, flash memory, optical data storage device, and also carrier waves (for example, transmission over the Internet). It also includes the implementation in the form of. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

According to the present invention described above, in order to output the input continuous grayscale image as a binarized image, the local region characteristics, that is, the local mean value, the flatness, and the edge intensity of the input image are calculated, and the localized adaptation is performed using these calculated characteristic values. By calculating the thresholds and comparing the calculated local adaptation thresholds with the pixels currently being processed, it is possible to highlight visually important boundary components while properly representing the background area,

Since the adaptive local threshold is determined by the distribution of the mask value corresponding to the pixel currently being processed and the current pixel surrounding value, even if a mask smaller than the size of the mask used in the conventional method is used, the regular pattern by the mask is not applied. Because it can be reduced, the hardware implementation has the effect of reducing the size of the memory for storing the threshold of the mask.

1 is a block diagram of a conventional image binarization apparatus

2 shows a conventional masking scheme.

FIG. 3 is a diagram showing an output image using a stokestick mask

4 shows a Bayer dither mask of 8 by 8 pixel size;

5 is a block diagram illustrating a binarization apparatus according to an embodiment of the present invention.

6 (a) is a diagram showing a current pixel and a peripheral region for explaining an embodiment of the present invention.

6 (b) is a block diagram showing a local windowing unit 510 used in the binarization apparatus according to the embodiment of the present invention shown in FIG.

FIG. 7 is a block diagram illustrating a local mean value detector 510 used in a binarization apparatus according to an embodiment of the present invention illustrated in FIG. 5.

FIG. 8 is a block diagram illustrating a mask threshold generator 550 used in the binarization apparatus according to the embodiment of the present invention shown in FIG. 5.

FIG. 9 is a block diagram illustrating a gamma error correcting unit 570 used in the binarization apparatus of FIG. 5 according to an embodiment of the present invention.

FIG. 10 is a block diagram illustrating a first weight calculator 580 used in the binarization apparatus shown in FIG. 5.

11 illustrates a parameter region used for functions for _obtaining F _Loc and F _MASK according to an embodiment of the present invention.

12 is a block diagram illustrating a first multiplier 600 and a second multiplier 610 used in the binarization apparatus shown in FIG. 5.

FIG. 13 is a block diagram illustrating a first adder 620 used in the binarization apparatus according to the embodiment of the present invention shown in FIG. 5.

FIG. 14 is a block diagram illustrating a second adder 630 used in the binarization apparatus according to the embodiment of the present invention shown in FIG. 5.

FIG. 15 is a block diagram illustrating a comparator 640 used in a binarization apparatus according to an embodiment of the present invention illustrated in FIG. 5.

Claims (5)

In the binarization method for outputting an input image as a binarized image, Calculating an adaptive local threshold value in consideration of the characteristics of the pixel peripheral area to be processed in the input image; And comparing the calculated adaptive local threshold with a value of the pixel to be currently processed to output a binary pixel value corresponding to the pixel to be currently processed. The method of claim 1, The adaptive local threshold calculation step is Calculating a local mean value representing a pixel mean value of peripheral pixel values of the pixel to be currently processed; Calculating flatness of the peripheral pixel based on the peripheral pixel values of the pixel to be currently processed; Modifying the calculated local mean value based on the calculated local mean value and flatness; Calculating an adaptive local threshold value using the modified local mean value. The method of claim 2, The adaptive local threshold calculation step is Calculating edge intensities of peripheral pixels based on the peripheral pixel values of the pixel to be currently processed; Generating a mask threshold value corresponding to the pixel to be currently processed; Modifying the generated mask threshold based on the calculated edge intensity, And wherein said adaptive local threshold calculation step uses said modified mask threshold. The method of claim 2, The adaptive local threshold calculation step is Calculating edge intensities of peripheral pixels based on the peripheral pixel values of the pixel to be currently processed; Generating a mask threshold value corresponding to the pixel to be currently processed; Modifying the generated mask threshold based on the calculated edge intensity and the calculated flatness, And wherein said adaptive local threshold calculation step uses said modified mask threshold. The method according to claim 3 or 4, The adaptive local threshold calculation step Calculating a weight for a local mean value based on the local mean value and flatness; Calculating a weight for a mask threshold using at least one of the flatness and the edge strength; The adaptive local threshold is calculated using the equation Here, Avg _Loc (x, y) is the local mean value, F _Loc is the weight value for the local mean value, M (x, y) is the mask threshold value, F _Mask is the weight value for the mask threshold value, Offset is an offset of the input image, F _Offset is a weighting method for gamma correction of the offset, characterized in that the image.