A picture recognition and replacement system

Abstract

The present invention relates to the technical field related to information processing, and discloses a picture recognition and replacement system. When screening and managing picture contents, after identifying the picture contents, illegal contents are replaced regionally, thereby changing the one-size-fits-all blocking method for the entire picture contents, and being able to effectively avoid the erroneous blocking of other important contents. Providing two different replacement methods can reduce the sense of disharmony of the picture contents after replacement and improve the quality of replacement.

Description

Technical Field

[0001] The present invention relates to the technical field related to information processing, and in particular to a picture recognition and replacement system. Background Art

[0002] With the rapid development of the Internet in recent years, the number of Internet users has increased significantly, and various group activities have gradually changed from offline to online. Therefore, the security management of network information is very important. Reasonable control of online publishing and dissemination of content can reduce the occurrence of crimes and other behaviors that are not conducive to social stability to a certain extent.

[0003] In the prior art, various screening management steps have been set up for the publication and transmission of illegal content on the Internet. Through the program, illegal information with bad content is screened and blocked, especially for text content, which can effectively filter it. For image information, illegal images with pornographic content can also be effectively identified and filtered.

[0004] However, the existing technology does not have a good way to deal with some transmission methods that mix text information with normal pictures, and the one-size-fits-all blocking method adopted in identification and filtering will not only cause misjudgment, but also often affect the transmission of some other important information and the normal work and life of relevant personnel (for example, a website intercepts the information sent by the user and adds seal content to it before disseminating it. Direct interception will cause the complete loss of the user's normal transmission content). Summary of the invention

[0005] The object of the present invention is to provide a picture recognition and replacement system to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A picture recognition and replacement system, comprising:

[0008] A text recognition module, used to process and recognize text on an image, generate a text recognition result, and judge the text recognition result according to a preset security judgment library, and output text replacement information according to the judgment result, wherein the text replacement information includes text position information corresponding to the text recognition result, the text replacement information is used to indicate that the text recognition result needs to be replaced, and the text position information is used to indicate the position of the pixel point of the text relative to the image;

[0009] A range color picking module, used to generate a replacement color picking range according to the text replacement information, and to pick colors for the image according to the replacement color picking range to generate a color sampling result, and to process and judge the color sampling result according to a preset replacement judgment standard and generate a judgment result, wherein the judgment result is used to guide the execution of a replacement color filling method or an overlay color filling method;

[0010] A replacement color filling module, used to execute a replacement color filling method, the replacement color filling method comprising: performing color analysis on the replacement color selection range to generate a color segmentation line, extending the color segmentation line to intersect the text, and filling and replacing the text according to the color segmentation line, the color segmentation line being used to represent the boundary position information of different colors;

[0011] An overlay coloring module is used to execute an overlay coloring method, the overlay coloring method comprising: performing color analysis on the replacement color range to generate a color gradient model, randomly generating replacement characters of the same amount of text as the text recognition result according to a preset replacement character library, generating and outputting an overlay layer according to the color gradient model and the replacement characters, the overlay layer being used to replace the content of the image at the text position information.

[0012] As a further solution of the present invention: the range color picking module specifically includes:

[0013] A text frame selection unit, used for reading the text position information in the text replacement information, and generating a rectangular text range according to the text position information, wherein the rectangular text range is a minimum rectangular range including all pixel points corresponding to the text;

[0014] A color picking frame selection unit, used to generate a replacement color picking range according to the rectangular text range and a preset color picking selection standard, wherein the replacement color picking range is a range area that imitates the rectangular text range;

[0015] A color sampling unit, used to sample the replacement color range of the image and generate a color sampling result;

[0016] The color determination unit is used to process and determine the color sampling result according to the preset replacement determination standard, and guide the execution of the replacement color filling method or the overlay color filling method according to the determination result.

[0017] As a further solution of the present invention: the color sampling unit specifically includes:

[0018] A grid generation subunit, used for performing a gridding separation process on the replacement color selection range to generate a plurality of color selection grids of uniform size, wherein the size of the color selection grid is determined by the replacement color selection range and a preset gridding separation process method;

[0019] The sampling and color picking subunit is used to pick colors from each of the color picking grids and generate color sampling results.

[0020] As a further solution of the present invention: the text recognition module specifically includes:

[0021] A preprocessing unit, used for preprocessing the image, wherein the preprocessing is used to make the image include only two different colors: text and background;

[0022] A text numbering unit, used to number the text content on the image, wherein the numbering marks are used to distinguish different content segments in different areas of the image;

[0023] A position recording unit, used to generate text position information according to the number mark;

[0024] A tilt correction unit, used for performing tilt correction on the text on the picture that is in a tilted state, wherein the tilt correction is used to make the directions of the text on the picture consistent;

[0025] A text recognition unit, used for recognizing text on the image after tilt correction processing, and generating a text recognition result, wherein the text recognition result corresponds to the text position information;

[0026] The text determination unit is used to determine the text recognition result according to the preset security determination library, generate a determination result, and if the determination result is a violation, generate and output text replacement information.

[0027] As a further solution of the present invention: the replacement color filling module specifically includes:

[0028] A color segmentation unit, configured to perform color analysis on the replacement color range to generate a color segmentation line, wherein the generated color segmentation line is discontinuous at the text;

[0029] A curvature simulation unit, used for simulating and generating a curvature change model of the color segmentation line;

[0030] A dividing line completion unit is used to complete and connect the two ends of the color dividing line at the text according to the curvature change model.

[0031] As a further solution of the present invention: the covering coloring module specifically includes:

[0032] A color analysis unit, used to obtain the color distribution within the replacement color range, and generate a color gradient model according to the color distribution analysis;

[0033] A replacement calculation unit, used for obtaining a replacement character quantity according to the character replacement information, wherein the replacement character quantity is used for representing the quantity of characters that need to be replaced;

[0034] A replacement generating unit, used for randomly generating replacement characters according to the amount of text and the preset replacement character library;

[0035] The text replacement unit is used to generate an overlay layer according to the color gradient model and the replacement characters, and output the overlay layer on the image.

[0036] As a further solution of the present invention: the replacement generation unit further includes a character scaling unit;

[0037] The character scaling unit is used to scale the replacement character according to the rectangular character range so that the rectangular character range is the minimum rectangular range of the replacement character.

[0038] Compared with the prior art, the beneficial effects of the present invention are: when screening and managing image content, after identifying the image content, the illegal content is replaced regionally, which changes the one-size-fits-all blocking method for the entire image content, can effectively avoid the erroneous blocking of other important content, and provides two different replacement methods to reduce the sense of incongruity of the image content after replacement and improve the quality of replacement. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A block diagram of an image recognition and replacement system.

[0040] Figure 2 A block diagram of the composition of a range color picking module in an image recognition replacement system.

[0041] Figure 3 A block diagram of the composition of a color sampling unit in an image recognition replacement system.

[0042] Figure 4 The present invention is a block diagram of the composition of a color determination unit in an image recognition replacement system. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0044] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0045] like Figure 1The image recognition and replacement system provided by one embodiment of the present invention comprises the following steps:

[0046] The text recognition module 100 is used to process and recognize text on an image, generate a text recognition result, and judge the text recognition result according to a preset security judgment library, and output text replacement information according to the judgment result. The text replacement information includes text position information corresponding to the text recognition result one by one. The text replacement information is used to indicate that the text recognition result needs to be replaced, and the text position information is used to indicate the position of the pixel point of the text relative to the image.

[0047] In this embodiment, the text recognition module is used to process the image, and then perform text recognition and security judgment on the processed image. Here, the text recognition result is the text content generated after the module recognizes the text on the image. The text position information corresponds to the text recognition result one by one, indicating the position on the image corresponding to a certain text recognition result. For example, there is a character "A" on a certain image. After being recognized by the text recognition module 100, a text recognition result with the content "A" and text position information corresponding to the text recognition result, which is the pixel position of the character "A", will be obtained. When there are multiple scattered characters on a certain image, a pair of corresponding text recognition results and corresponding text position information will be generated; the role of the security judgment library is to judge the text recognition results, that is, to play a role in network security supervision, to prevent certain groups from transmitting certain unsafe remarks or content by using online pictures.

[0048] The range color picking module 300 is used to generate a replacement color picking range according to the text replacement information, and to pick colors for the image according to the replacement color picking range to generate a color sampling result, and to process and judge the color sampling result according to a preset replacement judgment standard to generate a judgment result, and the judgment result is used to guide the execution of a replacement color filling method or an overlay color filling method.

[0049] In this embodiment, the replacement color range corresponds to and is larger than the text range on the image, and the boundary of the replacement color range does not intersect with the edge of the text and is spaced a certain distance apart (in order to further hide the text content in subsequent processing, the replacement color range here does not take a closed range that is similar to the text shape). The color sampling result refers to the number of colors within this replacement color range (not all color numbers can be represented in multiple color selection points), and the judgment result here refers to the distribution of colors within the color range. When there are too many or too few colors, different color filling methods will be guided to be executed, namely, the overlay color filling method and the replacement color filling method.

[0050] The replacement coloring module 500 is used to execute the replacement coloring method, which includes: performing color analysis on the replacement color range to generate a color dividing line, extending the color dividing line to intersect with the text, and filling and replacing the text according to the color dividing line, and the color dividing line is used to represent the boundary position information of different colors.

[0051] In this embodiment, the replacement coloring module 500 is used to execute the replacement coloring method. The replacement coloring method is used for the situation where the color of the picture with text background changes less. When in use, the color of the picture is analyzed in the selected range, the color is divided, and a color segmentation line is generated, which is similar to a color contour line, used to distinguish adjacent different colors. Then, the content of the text in the area divided by the color segmentation line is replaced with the corresponding color, thereby realizing the smearing of illegal or unsafe text content in the picture, which is convenient for replacement; here, because of the existence of text content, most of the color segmentation lines generated in the selected range are discontinuous (separation of text strokes), so it is necessary to extend the color segmentation line. When extending, the color segmentation line is generated by simulating the curvature change of the existing color segmentation line to connect the discontinuous parts (when extending a color segmentation line, the corresponding curvature change of the color segmentation line is adopted).

[0052] The overlay coloring module 700 is used to execute an overlay coloring method, which includes: performing color analysis on the replacement color range to generate a color gradient model, randomly generating replacement characters with the same amount of text as the text recognition result according to a preset replacement character library, and generating and outputting an overlay layer based on the color gradient model and the replacement characters, wherein the overlay layer is used to replace the content of the image located at the text position information.

[0053] In this embodiment, the overlay coloring method can be understood as generating a new layer based on the background color of the picture that can cover the text on the picture, and then covering the original text on the picture, thereby achieving covering the text in the picture that violates the rules or contains inappropriate remarks. The role of the color gradient model is to record the replacement of colors within the color selection range and the color changes, so as to imitate and generate a new layer, that is, the overlay layer.

[0054] like Figure 2 As shown, as another preferred embodiment of the present invention, the range color picking module 300 specifically includes:

[0055] The text frame selection unit 301 is used to read the text position information in the text replacement information, and generate a rectangular text range according to the text position information. The rectangular text range is a minimum rectangular range that includes all pixel points corresponding to the text.

[0056] The color picking frame selection unit 302 is used to generate a replacement color picking range according to the rectangular text range and a preset color picking selection standard, wherein the replacement color picking range is a range area that imitates the rectangular text range.

[0057] The color sampling unit 303 is used to sample the replacement color range of the image to generate a color sampling result.

[0058] The color determination unit 304 is used to process and determine the color sampling result according to the preset replacement determination standard, and guide the execution of the replacement color filling method or the overlay color filling method according to the determination result.

[0059] In this embodiment, the range color picking module 300 is further divided and the content execution is explained. When in use, the text is first framed to obtain a minimum rectangular range occupied by the text, and then the range is expanded through the color picking frame selection unit 302 to generate a replacement color picking range. The main use of the replacement color picking range is to increase the rectangular text range, thereby expanding the sample accuracy and persuasiveness when sampling and analyzing colors, and improving the representativeness of the sampling results in a larger range. However, at the same time, the entire content of the image is not used for analysis, because the colors of different areas of the image have different characteristics. Therefore, when the range is larger, the representativeness will be reduced and the processing amount of color content data will increase.

[0060] like Figure 3 As shown, as another preferred embodiment of the present invention, the color sampling unit 303 specifically includes:

[0061] The grid generation subunit 3031 is used to perform gridding and segmentation processing on the replacement color selection range to generate a plurality of color selection grids of uniform size, wherein the size of the color selection grid is determined by the replacement color selection range and a preset gridding and segmentation processing method.

[0062] The sampling and color picking subunit 1032 is used to pick colors from each of the color picking grids and generate color sampling results.

[0063] In this embodiment, when the color sampling unit 303 performs color sampling, it first needs to divide the area through the grid generation subunit 3031 to obtain several smaller blocks (i.e., color picking grids), and then use the sampling color picking subunit 1032 to pick colors for these blocks respectively. The area of ​​the color picking grid obtained by the division here depends on the user's setting. The smaller the area of ​​a single grid, the higher the color sampling accuracy. At the same time, the computing power consumption of the entire color picking process is also higher. Therefore, the size setting of the color picking grid is set by the user according to his own needs.

[0064] like Figure 4 As shown, as another preferred embodiment of the present invention, the color determination unit 304 specifically includes:

[0065] The color picking area generating subunit 3041 is used to calculate and generate the color picking range area according to the replacement color picking range and the text position information, and the color picking range area is the difference between the replacement color picking range area and the area occupied by the text.

[0066] The color density generating subunit 3042 is used to calculate and generate color density according to the color sampling result and the area of ​​the color sampling range, and the color density is used to represent the number of different colors existing in a unit area within the replacement color sampling range.

[0067] The color density determination subunit 3043 is used to determine the color density according to the preset replacement determination standard and generate a determination result, wherein the determination result includes a high-density color distribution and a low-density color distribution.

[0068] In this embodiment, the color judgment unit 304 is further decomposed and explained. In the process of making a judgment, the color judgment unit calculates the density of the color distribution, that is, the color density, and then makes a judgment based on the density. Of course, the replacement judgment standard here is a more reliable division standard obtained through machine learning adjustment, which is obtained through trial analysis of a large amount of data after the user sets the color grid.

[0069] As another preferred embodiment of the present invention, the range color picking module 300 further includes a replacement execution guiding unit:

[0070] The replacement execution guiding unit is used to respond to the determination result, and when the determination result is a low-density color distribution, guide the execution of a replacement coloring method, and when the determination result is a high-density color distribution, guide the execution of a covering coloring method.

[0071] In this embodiment, the replacement execution guiding unit is a guiding judgment unit for program jump, which is used to guide the replacement coloring module 500 and the overlay replacement module 700 to work respectively, so as to perform the next step of image and text processing.

[0072] As another preferred embodiment of the present invention, the text recognition module 100 specifically includes:

[0073] The preprocessing unit is used to preprocess the image, and the preprocessing is used to make the image only include two different colors: text and background.

[0074] The text numbering unit is used to number the text content on the image, and the numbering marks are used to distinguish different content segments in different areas on the image.

[0075] A position recording unit is used to generate text position information according to the number mark.

[0076] The tilt correction unit is used to perform tilt correction on the text on the picture that is in a tilted state, and the tilt correction is used to make the directions of the text on the picture consistent.

[0077] The text recognition unit is used to recognize the text on the image after the tilt correction process and generate a text recognition result, wherein the text recognition result corresponds to the text position information.

[0078] The text determination unit is used to determine the text recognition result according to the preset security determination library, generate a determination result, and if the determination result is a violation, generate and output text replacement information.

[0079] In this embodiment, the text recognition module 100 is explained here. The text recognition method is mainly performed through OCR and other technologies. The difference compared with the prior art is that the text recognition module 100 includes a text numbering unit and a position recording unit. Multiple text segments may be included in the same picture. For example, in a certain picture, a section of text is horizontally arranged at the upper position, and a section of text is inclined at the bottom position. The two sections of text correspond to two numbers for distinction, and then each number will correspond to a set of text position information, which is used to record the original position of the text, that is, the original pixel occupancy range, for the convenience of subsequent replacement work.

[0080] As another preferred embodiment of the present invention, the replacement color filling module 500 specifically includes:

[0081] The color segmentation unit is used to perform color analysis on the replacement color range to generate a color segmentation line, and the generated color segmentation line is discontinuous at the text.

[0082] The curvature simulation unit is used to simulate and generate a curvature change model of the color segmentation line.

[0083] A dividing line completion unit is used to complete and connect the two ends of the color dividing line at the text according to the curvature change model.

[0084] In this embodiment, the replacement color filling module 500 is explained. The main execution method is to divide the color using a color dividing line, and then simulate the curvature of the dividing line to complete the dividing line that is broken at the text. In this way, the color between two adjacent color dividing lines is the same, and the text that needs to be replaced can be smeared and eliminated, thereby achieving the purpose.

[0085] As another preferred embodiment of the present invention, the overlay coloring module 700 specifically includes:

[0086] The color analysis unit is used to obtain the color distribution within the replacement color range, and generate a color gradient model based on the color distribution analysis.

[0087] The replacement calculation unit is used to obtain the replacement character quantity according to the character replacement information, and the replacement character quantity is used to represent the number of characters that need to be replaced.

[0088] The replacement generating unit is used to randomly generate replacement characters according to the amount of text and the preset replacement character library.

[0089] The text replacement unit is used to generate an overlay layer according to the color gradient model and the replacement characters, and output the overlay layer on the image.

[0090] As another preferred embodiment of the present invention, the replacement generation unit further includes a character scaling unit;

[0091] The character scaling unit is used to scale the replacement character according to the rectangular character range so that the rectangular character range is the minimum rectangular range of the replacement character.

[0092] In this embodiment, the overlay coloring module 700 is explained here. This step is to generate a small-range base map according to the color distribution within the range. The color on the base map is simulated according to the collected color distribution. Therefore, it has a certain similarity with the corresponding position of the original image. After the overlay replacement, the degree of disharmony can be reduced. Then, replacement characters are added to the base map to overlay and replace the text content of the original image to achieve the purpose.

[0093] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0094] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the disclosure in the specification and examples. This application is intended to cover any variations, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

[0095] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A picture recognition and replacement system, characterized in that: Include: A text recognition module, used to process and recognize text on an image, generate a text recognition result, and judge the text recognition result according to a preset security judgment library, and output text replacement information according to the judgment result, wherein the text replacement information includes text position information corresponding to the text recognition result, the text replacement information is used to indicate that the text recognition result needs to be replaced, and the text position information is used to indicate the position of the pixel point of the text relative to the image; A range color picking module, used to generate a replacement color picking range according to the text replacement information, and to pick colors for the image according to the replacement color picking range to generate a color sampling result, and to process and judge the color sampling result according to a preset replacement judgment standard and generate a judgment result, wherein the judgment result is used to guide the execution of a replacement color filling method or an overlay color filling method; A replacement color filling module, used to execute a replacement color filling method, the replacement color filling method comprising: performing color analysis on the replacement color selection range to generate a color segmentation line, extending the color segmentation line to intersect the text, and filling and replacing the text according to the color segmentation line, the color segmentation line being used to represent the boundary position information of different colors; A covering color filling module is used to execute a covering color filling method, wherein the covering color filling method comprises: performing color analysis on the replacement color selection range to generate a color gradient model, randomly generating replacement characters of the same amount of text as the text recognition result according to a preset replacement character library, generating and outputting a covering layer according to the color gradient model and the replacement characters, wherein the covering layer is used to replace the content of the picture at the text position information; The range color picking module specifically includes: A text frame selection unit, used for reading the text position information in the text replacement information, and generating a rectangular text range according to the text position information, wherein the rectangular text range is a minimum rectangular range including all pixel points corresponding to the text; A color picking frame selection unit, used to generate a replacement color picking range according to the rectangular text range and a preset color picking selection standard, wherein the replacement color picking range is a range area that imitates the rectangular text range; A color sampling unit, used to sample the replacement color range of the image and generate a color sampling result; A color determination unit, used for processing and determining the color sampling result according to the preset replacement determination standard, and guiding the execution of the replacement color filling method or the overlay color filling method according to the determination result; The color sampling unit comprises: A grid generation subunit, used for performing a gridding separation process on the replacement color selection range to generate a plurality of color selection grids of uniform size, wherein the size of the color selection grid is determined by the replacement color selection range and a preset gridding separation process method; A sampling color picking subunit, used to pick colors from each of the color picking grids and generate color sampling results; The color determination unit comprises: A color picking area generating subunit, used for calculating and generating a color picking range area according to the replacement color picking range and the text position information, wherein the color picking range area is the difference between the replacement color picking range area and the area occupied by the text; A color density generating subunit, used for calculating and generating color density according to the color sampling result and the area of ​​the color sampling range, wherein the color density is used to characterize the number of different colors existing in a unit area within the replacement color sampling range; A color density determination subunit, configured to determine the color density according to the preset replacement determination standard and generate a determination result, wherein the determination result includes a high-density color distribution and a low-density color distribution; The range color picking module also includes a replacement execution guiding unit; The replacement execution guiding unit is used to respond to the determination result, and when the determination result is a low-density color distribution, guide the execution of a replacement coloring method, and when the determination result is a high-density color distribution, guide the execution of a covering coloring method.

2. The image recognition and replacement system according to claim 1, characterized in that: The text recognition module comprises: A preprocessing unit, used for preprocessing the image, wherein the preprocessing is used to make the image include only two different colors: text and background; A text numbering unit, used to number the text content on the image, wherein the numbering marks are used to distinguish different content segments in different areas of the image; A position recording unit, used to generate text position information according to the number mark; A tilt correction unit, used for performing tilt correction on the text on the picture that is in a tilted state, wherein the tilt correction is used to make the directions of the text on the picture consistent; A text recognition unit, used for recognizing text on the image after tilt correction processing, and generating a text recognition result, wherein the text recognition result corresponds to the text position information; The text determination unit is used to determine the text recognition result according to the preset security determination library, generate a determination result, and if the determination result is a violation, generate and output text replacement information.

3. The image recognition and replacement system according to claim 1, characterized in that: The replacement color filling module includes: A color segmentation unit, used for performing color analysis on the replacement color range to generate a color segmentation line, wherein the generated color segmentation line is discontinuous at the text; A curvature simulation unit, used for simulating and generating a curvature change model of the color segmentation line; A dividing line completion unit is used to complete and connect the two ends of the color dividing line at the text according to the curvature change model.

4. The image recognition and replacement system according to claim 1, characterized in that: The covering coloring module includes: A color analysis unit, used to obtain the color distribution within the replacement color range, and generate a color gradient model according to the color distribution analysis; A replacement calculation unit, used for obtaining a replacement character quantity according to the character replacement information, wherein the replacement character quantity is used for representing the quantity of characters that need to be replaced; A replacement generating unit, used for randomly generating replacement characters according to the replacement text amount and a preset replacement character library; The text replacement unit is used to generate an overlay layer according to the color gradient model and the replacement characters, and output the overlay layer on the image.

5. The image recognition and replacement system according to claim 4, characterized in that: The replacement generation unit also includes a character scaling unit; The character scaling unit is used to scale the replacement character according to the rectangular character range so that the rectangular character range is the minimum rectangular range of the replacement character.

Images