CN112215781B - Improved local binarization method - Google Patents
Improved local binarization method Download PDFInfo
- Publication number
- CN112215781B CN112215781B CN202011180976.3A CN202011180976A CN112215781B CN 112215781 B CN112215781 B CN 112215781B CN 202011180976 A CN202011180976 A CN 202011180976A CN 112215781 B CN112215781 B CN 112215781B
- Authority
- CN
- China
- Prior art keywords
- color
- gray
- picture
- colors
- binarization method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 239000003086 colorant Substances 0.000 claims abstract description 33
- 238000012163 sequencing technique Methods 0.000 claims abstract description 4
- 238000004364 calculation method Methods 0.000 claims description 16
- 238000003491 array Methods 0.000 claims description 4
- 239000000126 substance Substances 0.000 claims description 3
- 238000000605 extraction Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 229910002056 binary alloy Inorganic materials 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
Images
Classifications
-
- G06T5/94—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration by the use of histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
Abstract
The invention discloses an improved local binarization method, which comprises the following steps: converting an input picture into a gray-scale image, and calculating the gray-scale value of each pixel point in the gray-scale image; counting peaks appearing in the gray level histogram and gray level colors corresponding to the peaks; reducing the gray color into RGB three-channel color, and combining similar colors; counting the number of pixel points of each color in the picture, calculating the occupation ratio, sequencing, and outputting a final color dictionary and the occupation ratio; selecting a plurality of groups of colors with the largest color ratio from the color dictionary in the step (5); acquiring an original design template picture of an input picture, and judging whether each color corresponds to a foreground color or a background color in the template picture; and replacing the colors corresponding to the foreground color and the background color with white and black in the binary image. The method can effectively improve the capability of correctly separating the foreground from the background in the picture with the complicated color background.
Description
Technical Field
The invention relates to the technical field of computer vision and graphic processing, in particular to an improved local binarization method.
Background
The image binarization method is one of the simplest image processing methods in the field of computer vision and graphic processing, is used as the basis of some complex image processing, and takes data after image binarization as input data to play the roles of simplifying data and improving the calculation speed of subsequent image processing. Whether the image binarization precision meets the requirement or not can lead to the subsequent judgment of certain complex image processing, and finally lead to the error of the result. Besides simplifying calculation, the image binarization can also separate an interested area from a background to distinguish a foreground from the background, and is widely applied to text document processing.
The basic idea of the traditional image binarization method needs to define a color threshold, the color threshold is determined by calculating a gray level histogram of an image and utilizing different color threshold selection methods, and then the color value of each pixel point in the image is judged, the white point is larger than the threshold, and the black point is smaller than or equal to the threshold. For example, patent specification No. CN100363942C discloses a binarization method for images, which comprises the following steps: (1) converting an input document into digital image data; (2) firstly, partitioning an image into a whole level, a sub-image area level and a pixel area level according to three levels, and then scanning and counting characteristic values of blocks in the three levels; the whole level is that the whole picture is a block, the regional level of the sub-picture is that the picture is divided into several sub-pictures, the said sub-picture is the fixed size, or according to the size of the whole level picture, confirm in proportion, every sub-picture block is not smaller than 128 x 128 pixel points; the pixel region level takes a pixel lattice of n x n as a block, wherein n is a positive integer and is less than or equal to 16; (3) calculating the threshold value of each block of the whole level and the sub-image regional level; (4) and (4) correcting the characteristic value of the sub-image region level according to the data obtained in the step (2) and the step (3). This method is very harsh on the selection of the color threshold, which directly affects the binarization result.
In addition, other common binarization algorithms can be roughly classified into six categories: empirical value based methods, global threshold based methods, local threshold based methods, iterative based methods, maximum entropy based methods, maximum inter-class variance (OSTU). The methods mainly use characteristics such as double peaks and multiple peaks in a gray level histogram to determine a color threshold value, so that a relatively good binarization result can be obtained, but when certain specific conditions are met, for example, in character extraction under a complex color background, characteristics of double peaks and multiple peaks in the gray level histogram are not obvious, color distribution is relatively discrete, a foreground and a background cannot be separated correctly, so that an image after binarization is too dark or too light in color, and binarization failure is finally caused.
In the process of enterprise Logo in daily life, the Logo usually appears in a background image with disordered colors, the approximate position of the Logo can be directly detected by using the existing machine learning method, but an image area close to the Logo cannot be obtained, and the method cannot achieve a good extraction effect, so that the invention aims to provide a method for separating the foreground from the background of a picture with a complex color background by means of an original design image template.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides an improved local binarization method which can effectively improve the capability of correctly separating the foreground from the background in the picture with the complicated color background.
An improved local binarization method comprises the following steps:
(1) converting an input picture into a gray-scale image, and calculating the gray-scale value of each pixel point in the gray-scale image;
(2) counting peaks appearing in the gray level histogram and gray level colors corresponding to the peaks;
(3) reducing the gray color into RGB three-channel color, and combining similar colors;
(4) counting the number of pixel points of each color in the picture after combination in the step (3), calculating the proportion of each color, sequencing, and outputting a final color dictionary and the proportion;
(5) selecting a plurality of groups of colors with the largest color ratio from the color dictionary in the step (4);
(6) acquiring an original design template picture of an input picture, traversing each color selected in the step (5), and judging whether each color corresponds to a foreground color or a background color in the template picture;
(7) and (4) replacing the color corresponding to the foreground color in the step (6) with white in the binary image, and replacing the color corresponding to the background color with black in the binary image to obtain the final binary image.
By adopting the method, the foreground can be extracted from the picture with complex background color, the influence of the threshold value in the traditional binarization algorithm is avoided, and the binarization robustness is improved.
Preferably, in the step (1), a calculation formula of the gray value of each pixel point in the gray map is as follows:
wherein, PgrayRepresenting the gray value of the current pixel point, round () representing rounding, and R, G and B are the color values of three channels of R, G and B of each pixel point.
Each channel value divided by 2NThe reason for (2) is to reduce the complexity in the calculation2×(8-N)And 28-NThe shift can be made such that each gray value can contain RGB three-channel information.
Further preferably, N is 3; the calculation formula of the gray value of the current pixel point is as follows:
thus, the picture has a gray scale level of 0-23×(8-3)I.e. 0-32767, total 23×(8-3)I.e., 32767 color levels.
R, G, B the value ranges of the three channels are 0-255 decimal, and the range of the second power is 0-28If a number is converted to binary, 8 bits will correspond. For example, decimal 132 is converted to binary is (10000100)2。
The division by 8 is intended to define the number of bits of the resulting color value, so that each channel occupies 5 bits in the binary system, which makes it possible to distribute the color values between 0 and 32767, with the corresponding second power ranging from 0 to 215。
The purpose of multiplying 1024 is to shift the value of the R channel left 10 bits to 15-10 bits, and the purpose of multiplying 32 is to be able to shift the G channel color left 5 bits, occupying the middle 5 bits, and the B channel color occupying the last 5 bits, where each block occupies 5 bits.
Preferably, before the statistics in the step (2), the sum of the numbers of the pixels in the gray histogram is normalized to be not more than 1000000. The size of the partial image is typically 1000 × 1000, here set to 1000000, which on the one hand can satisfy the size requirements and on the other hand can reduce the time loss in the calculation.
Preferably, the peak value in step (2) is not less than 20000, and the minimum interval between two peaks is 1. As described above, the total pixel point is normalized to 1000000, and the peak value in step (2) is set to 20000 in order to remove the influence of the color with a very low ratio on the binarization. The empirical value is set to 2% of the total pixel points, and the specific embodiment can be dynamically modified.
Preferably, in the step (2), before counting peaks appearing in the gray histogram, zero data is added to the head and the tail of the gray histogram.
In step (2), there may be peaks at the first position and the last position of the histogram, because in the grayscale histogram, the RGB color corresponding to the first position is black, and the RGB color corresponding to the last position is white. In a typical text recognition scenario, white and black are more likely to appear, and therefore, zero data needs to be added to the head and tail of the gray histogram to ensure that the peak is correctly found when searching for the peak.
Preferably, the step (3) is specifically: converting the gray value of the gray color into a reduced RGB three-channel numerical value, defining a minimum color interval number, if the interval difference of the reduced color in the same channel does not exceed the interval number, determining the color as a similar color, and combining the color proportions of the similar colors. In the process of color change, in order to improve the calculation efficiency, a bit operation method is adopted for shifting to carry out calculation, and the calculation speed is improved.
Preferably, the step (6) includes:
(61) defining color arrays
Obtaining the color array of each color selected in the step (5)
(62) Obtaining an original design template picture of an input picture, and obtaining a color array of a foreground color of the template picture
(63) Computing
And
if the correlation degree is larger than the threshold lambda, the color is considered to be white corresponding to the template picture, and a calculation formula of the correlation degree is as follows:
wherein the content of the first and second substances,
is composed of
And
the covariance of (a) of (b),
is composed of
The variance of (a) is calculated,
is composed of
The variance of (c).
The absolute value of the correlation coefficient r is 0.8 or more, and a strong correlation is considered between the two, a weak correlation is considered between 0.3 and 0.8, and no correlation is considered between 0.3 and less.
The invention has the beneficial effects that:
by shifting the RGB three-channel colors, each color interval is artificially expanded, the diversity of the colors can be better reflected, the subsequent separation of foreground colors is facilitated, meanwhile, the method is not influenced by a threshold value in the traditional binarization algorithm, the binarization robustness is increased, and the capability of correctly separating the foreground from the background in a picture with a complex color background is improved.
Drawings
Fig. 1 is a picture (actually, a color picture) to be subjected to binarization operation;
fig. 2 is a template picture of an input picture adopted in the present embodiment;
fig. 3 is a picture after the binarization operation.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
An improved local binarization method, an original picture needing binarization operation is shown in fig. 1 (actually, a color picture), and includes the following steps:
(1) converting an input picture (fig. 1) into a grayscale map and counting a grayscale histogram
The gray value calculation formula of each pixel point in the gray image is as follows:
wherein, PgrayRepresenting the gray value of the current pixel point, round () representing rounding, and R, G, B being the colors of three channels of R (red), G (green), B (blue) of each pixel pointNumerical values.
In this embodiment, N is defined as 3, so the gray value calculation formula of each pixel in the gray map is as follows:
thus, the picture has a gray scale level of 0-23×(8-3)I.e. 0-32767, total 23×(8-3)And 32767 color levels.
(2) Normalizing the sum of the number of the pixel points of the gray histogram to a specified size, and counting peaks appearing in the gray histogram and gray colors corresponding to the peaks
In this embodiment, the sum of the number of pixels in the histogram is normalized to 1000000 pixels; the gray histogram is a one-bit array with length of 32768 and is named as Color, wherein each element is named as Colori. Each array subscript i represents one of 0-32767 gray colors, and the data in the array table represents the number of pixels of the gray color after normalization; such as Color327672000 means that the gray color is 32767, and a total of 2000 pixels are the gray color in the present normalized to 1000000 pixels.
And counting peaks appearing in the gray level histogram by using a function for searching the peak values, wherein the minimum height of the peak values is set to be 20000 according to set parameters, and the minimum interval between the two peaks is set to be 1. There may be a peak at the first position and the last position of the histogram in this step, and a zero data needs to be added at the head and tail of the gray histogram respectively to ensure that the peak can be correctly found when searching for the peak.
(3) Reducing the gray color into RGB three-channel color, and combining similar colors;
in this embodiment, the original gray value is converted into an RGB three-channel numerical value reduced by 8 times, the minimum color interval number is defined as 1, if the interval difference between the reduced colors in the same channel is 1, the reduced colors are considered to be similar colors, and the two colors are combined according to the color ratio; that is, when the absolute value of the difference between the R channel, B channel, and G channel values is 8 or less, the two gray colors are determined to be similar colors.
(4) Counting the number of pixel points of each color in the picture after combination in the step (3), calculating the proportion of each color, sequencing, and outputting a final color dictionary and the proportion;
(5) selecting a plurality of groups of colors with the largest color ratio from the color dictionary in the step (4);
in this embodiment, the number of colors is greater than 4, and the 3 colors having the largest ratio are selected.
(6) Acquiring an original design template picture of an input picture, traversing each color selected in the step (5), and judging whether each color corresponds to a foreground color or a background color in the template picture;
define the color array of color as
If the picture size is width × height, the array is a width × 1 array, and there are total width elements, and each element represents the total number of pixels with color in the vertical direction.
The picture to be subjected to binarization processing is shown in fig. 1, the size of the picture is 759 × 147, the number of colors to be compared obtained in step (5) is 3, the three color arrays are arranged from large to small according to the proportion of the image, and the three color arrays are respectively recorded as 759 × 1
And
the method comprises the steps of obtaining an original design template picture of an input picture, wherein the template picture can be directly obtained, as shown in fig. 2, the template picture has a black background, a (0, 0, 0) RGB color, a white foreground, a (255, 255, 255) RGB color and no other colors, and is divided into two areas, one area is an icon area (corresponding to the area in fig. 2b), and the other area is a text area (corresponding to the area in fig. 2C).
Wherein, the template picture comprises three sub-templates, which are respectively marked as template (fig. 2a), template (fig. 2b) and template (fig. 2c), the three sub-template pictures are scaled to the picture size to be binarized, i.e. 759 × 147, the foreground color in the template picture is white, the corresponding RBG is (255, 255, 255), the background color is black, the corresponding RGB is (0, 0, 0), then a white color array with the size of 759 × 1 can be obtained, which are respectively marked as template
And
calculating the correlation between the color array of the three colors obtained from the input picture (fig. 1) and the color array of the white foreground color in the template picture, and setting a threshold value of the correlation to be λ, in this embodiment, setting the threshold value of the correlation to be 0.4, and if the threshold value is greater than λ, considering that the color is white in the corresponding template picture.
The correlation calculation formula is as follows:
wherein the content of the first and second substances,
is composed of
And with
The covariance of (a) is determined,
is composed of
The variance of (a) is determined,
is composed of
The variance of (c).
The absolute value of the correlation coefficient r is 0.8 or more, and a strong correlation is considered between the two, a weak correlation is considered between 0.3 and 0.8, and no correlation is considered between 0.3 and less.
In the present embodiment of the present invention,
the threshold value is satisfied and the threshold value is satisfied,
if the threshold is satisfied, it is determined that the color (0, 0, 0) larger than the first corresponds to a text color, which corresponds to white in the template picture, i.e., the foreground color, and the color (183, 187, 190) larger than the second corresponds to an icon color, which corresponds to white in the template picture, i.e., the foreground color, and the remaining colors are black, i.e., the background color.
(7) And (4) replacing the color (character color + icon color) corresponding to the foreground color judged in the step (6) with white in the binary image, and replacing the color corresponding to the background color with black in the binary image to obtain the final binary image, as shown in fig. 3.
Although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that various changes in the embodiments and/or modifications of the invention can be made, and equivalents and modifications of some features of the invention can be made without departing from the spirit and scope of the invention.
Claims (7)
1. An improved local binarization method is characterized in that: the method comprises the following steps:
(1) converting an input picture into a gray-scale image, and calculating the gray-scale value of each pixel point in the gray-scale image;
(2) counting peaks appearing in the gray level histogram and gray level colors corresponding to the peaks;
(3) reducing the gray color into RGB three-channel color, and combining similar colors;
(4) counting the number of pixel points of each color in the picture after combination in the step (3), calculating the proportion of each color, sequencing, and outputting a final color dictionary and the proportion;
(5) selecting a plurality of groups of colors with the largest color ratio from the color dictionary in the step (4);
(6) acquiring an original design template picture of an input picture, traversing each color selected in the step (5), and judging whether each color corresponds to a foreground color or a background color in the template picture;
(7) replacing the color corresponding to the foreground color in the step (6) with white in the binary image, and replacing the color corresponding to the background color with black in the binary image to obtain a final binary image;
the step (6) comprises:
(61) defining color arrays
Obtaining the color array of each color selected in the step (5)
(62) Obtaining a color array of a foreground color of an original design template picture
(73) Computing
And
if the correlation degree is larger than the threshold lambda, the color is considered to be white corresponding to the template picture, and a calculation formula of the correlation degree is as follows:
wherein the content of the first and second substances,
is composed of
And
the covariance of (a) is determined,
is composed of
The variance of (a) is determined,
is composed of
The variance of (c).
2. The improved local binarization method as defined in claim 1, characterized in that: in the step (1), a calculation formula of the gray value of each pixel point in the gray map is as follows:
wherein, PgrayRepresenting the gray value of the pixel point, rounding () representing rounding, R, G and B are the color values of three channels of R, G and B of each pixel point, and N is the precision in calculation.
3. The improved local binarization method as recited in claim 2, characterized in that: and N is 3.
4. The improved local binarization method as defined in claim 1, characterized in that: before the statistics in the step (2), the sum of the number of the pixels of the gray histogram is normalized to be not more than 1000000.
5. The improved local binarization method as recited in claim 4, characterized in that: the peak value of the peak in the step (2) is not less than 20000.
6. The improved local binarization method as defined in claim 1, characterized in that: and (2) adding zero data to the head and the tail of the gray level histogram respectively before counting peaks appearing in the gray level histogram.
7. The improved local binarization method as recited in claim 1, characterized in that: the step (3) is specifically as follows: converting the gray value of the gray color into a reduced RGB three-channel numerical value, defining a minimum color interval number, if the interval difference of the reduced color in the same channel does not exceed the interval number, determining the color as a similar color, and combining the color proportions of the similar colors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011180976.3A CN112215781B (en) | 2020-10-29 | 2020-10-29 | Improved local binarization method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011180976.3A CN112215781B (en) | 2020-10-29 | 2020-10-29 | Improved local binarization method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112215781A CN112215781A (en) | 2021-01-12 |
| CN112215781B true CN112215781B (en) | 2022-07-12 |
Family
ID=74057506
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011180976.3A Active CN112215781B (en) | 2020-10-29 | 2020-10-29 | Improved local binarization method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112215781B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115797494B (en) * | 2023-02-06 | 2023-05-23 | 武汉精臣智慧标识科技有限公司 | Picture binarization processing method and device, electronic equipment and storage medium |
| CN116303909B (en) * | 2023-04-26 | 2023-08-22 | 山东齐鲁电子招标采购服务有限公司 | Matching method, equipment and medium for electronic bidding documents and clauses |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101571419A (en) * | 2009-06-15 | 2009-11-04 | 浙江大学 | Method adopting image segmentation for automatically testing LED indicator light of automobile instruments |
| CN106529543A (en) * | 2016-11-02 | 2017-03-22 | 徐庆 | Method and system for dynamically calculating multi-color-grade binary adaptive threshold |
| CN109189678A (en) * | 2018-08-22 | 2019-01-11 | 中国平安人寿保险股份有限公司 | A kind of webpage function verification method, computer readable storage medium and server |
| JP2019092101A (en) * | 2017-11-16 | 2019-06-13 | キヤノン株式会社 | Image processing apparatus, image processing method, and program |
| CN111242888A (en) * | 2019-12-03 | 2020-06-05 | 中国人民解放军海军航空大学 | Image processing method and system based on machine vision |
-
2020
- 2020-10-29 CN CN202011180976.3A patent/CN112215781B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101571419A (en) * | 2009-06-15 | 2009-11-04 | 浙江大学 | Method adopting image segmentation for automatically testing LED indicator light of automobile instruments |
| CN106529543A (en) * | 2016-11-02 | 2017-03-22 | 徐庆 | Method and system for dynamically calculating multi-color-grade binary adaptive threshold |
| JP2019092101A (en) * | 2017-11-16 | 2019-06-13 | キヤノン株式会社 | Image processing apparatus, image processing method, and program |
| CN109189678A (en) * | 2018-08-22 | 2019-01-11 | 中国平安人寿保险股份有限公司 | A kind of webpage function verification method, computer readable storage medium and server |
| CN111242888A (en) * | 2019-12-03 | 2020-06-05 | 中国人民解放军海军航空大学 | Image processing method and system based on machine vision |
Non-Patent Citations (1)
| Title |
|---|
| 《Fast Logo Detection and Recognition in Document Images》;Zhe Li 等；;《2010 International Conference on Pattern Recognition》;20101231;第2716-2719页； * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112215781A (en) | 2021-01-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10455117B2 (en) | Image processing apparatus, method, and storage medium | |
| EP1014699B1 (en) | Method of selecting colors for pixels within blocks for block truncation encoding | |
| US7577311B2 (en) | Color fringe desaturation for electronic imagers | |
| JP4016342B2 (en) | Apparatus and method for code recognition | |
| US7116443B2 (en) | Neutral pixel detection using color space feature vectors wherein one color space coordinate represents lightness | |
| CN112215781B (en) | Improved local binarization method | |
| US6078410A (en) | Image processing apparatus | |
| US8000535B2 (en) | Methods and systems for refining text segmentation results | |
| US7046390B2 (en) | System and method for scaling and enhancing color text images | |
| CN111372080A (en) | Processing method and device of radar situation map, storage medium and processor | |
| US6259813B1 (en) | Method and apparatus for image processing using a signal of multiple scale values | |
| US20040001230A1 (en) | Image processing apparatus and method thereof | |
| JP5151708B2 (en) | Image processing apparatus and image processing program | |
| US6744532B1 (en) | Method for enhancing printing quality | |
| JP4695914B2 (en) | Image processing apparatus and image processing method | |
| CN110807747B (en) | Document image noise reduction method based on foreground mask | |
| CN111402189B (en) | Video image color cast detection device and method | |
| US20080253655A1 (en) | Text and graphic separation method and text enhancement method | |
| JPH0654180A (en) | Method and device for image area separation | |
| CN111491069B (en) | Automatic setting method for color mode of document image | |
| Li et al. | A new color-to-gray conversion method based on edge detection | |
| CN111695557B (en) | Image processing method and device | |
| JP2008227825A (en) | Image processor, image forming apparatus provided with the same, and image processing method | |
| CN109712126B (en) | Picture identification method and device | |
| JP2005020413A (en) | Image processor, image processing method and program for making computer execute the method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |