Precise Identification Method and System for Sealing Area Based on Differential Dilated Convolution Verification

Abstract

The present invention discloses a method and system for accurately identifying a stamped area based on differential dilated convolution verification, which relates to the technical field of image recognition. The method includes the following steps: selecting an image template of the stamped area; inputting an image of the document to be stamped, calculating using a hash coding algorithm, and obtaining potential stamped areas in the image of the document to be stamped according to the similarity between the image of the document to be stamped and the image template of the stamped area; filtering the potential stamped areas using convolution kernels with different dilation rates respectively, and calculating the entropy values of the filtered potential stamped areas to obtain multiple area entropy values; comparing each area entropy value with a preset entropy value threshold to generate multiple comparison results; performing statistics on each comparison result, generating and determining the target stamped area according to the statistical result. The present invention combines differential dilated convolution filtering and entropy value verification, improving the accuracy of stamped area recognition.

Description

Technical Field

[0001] The present invention relates to the field of image recognition technology, and in particular to a method and system for accurately identifying stamped areas based on differential hole convolution verification. Background Art

[0002] With technological advancements, machine stamping has increasingly replaced manual stamping and is widely used in various fields, including business and finance, playing an increasingly important role. Machine stamping not only significantly reduces labor consumption but also greatly improves stamping efficiency. Furthermore, compared to manual stamping, machine stamping requires identifying the target stamping area within the document to be stamped, allowing it to be accurately stamped.

[0003] However, traditional methods still have certain errors when identifying the target stamping area, which may occasionally result in stamping in non-target areas. This can cause certain economic losses and even legal liabilities. Therefore, developing a method for accurately identifying the stamping area is a very important and urgent task. Summary of the Invention

[0004] In order to overcome the above problems or at least partially solve the above problems, an embodiment of the present invention provides a method and system for accurately identifying stamped areas based on differential void convolution verification, which combines differential void convolution filtering and entropy value verification to improve the accuracy of stamped area recognition.

[0005] The embodiment of the present invention is achieved as follows:

[0006] In a first aspect, an embodiment of the present invention provides a method for accurately identifying a stamped area based on differential hole convolution verification, comprising the following steps:

[0007] Select the stamp area image template;

[0008] Input the image of the document to be stamped, calculate using the hash coding algorithm and obtain the potential area to be stamped in the image of the document to be stamped according to the similarity between the image of the document to be stamped and the stamping area image template;

[0009] The potential areas to be stamped are filtered using convolution kernels with different void ratios, and the entropy values of the filtered potential areas to be stamped are calculated to obtain multiple area entropy values;

[0010] Compare the entropy value of each region with the preset entropy value threshold to generate multiple comparison results;

[0011] The comparison results are statistically analyzed, and the target stamping area is generated and determined based on the statistical results.

[0012] In order to solve the technical problems of certain errors and low recognition accuracy in the process of identifying the target stamping area in the prior art, this method combines the methods of differential dilated convolution filtering and entropy value verification to accurately identify the stamping area, greatly improving the recognition accuracy for subsequent precise stamping. First, a large number of stamping area image templates are selected to provide reference data for the follow-up; then, the scanned image of the document to be stamped is input, and the hash coding algorithm is used to find the potential stamping areas in the document to be stamped with high similarity to the area image template; then, the potential stamping areas are filtered respectively using convolution kernels with different dilation rates, and the entropy value of this area is calculated after filtering. Then, the entropy values of each area are compared with the preset entropy value threshold to generate multiple comparison results, and statistics are performed on each comparison result to generate and determine the target stamping area according to the statistical results.

[0013] This method uses the combined method of differential dilated convolution filtering and entropy value verification to improve the accuracy of stamping area recognition. It does not use relatively complex methods or a large amount of training data, reducing the calculation time and resource consumption.

[0014] Based on the first aspect, in some embodiments of the present invention, the method for selecting the stamping area image template includes the following steps:

[0015] Obtain multiple initial stamping area image template samples and establish a template database;

[0016] Use the saliency detection method to screen the initial stamping area image templates in the template database to obtain the stamping area image templates.

[0017] Based on the first aspect, in some embodiments of the present invention, the method for using the saliency detection method to screen the initial stamping area image templates in the template database to obtain the stamping area image templates includes the following steps:

[0018] Use the ST model to perform saliency detection on the initial stamping area image templates in the template database to screen and obtain the stamping area image templates.

[0019] Based on the first aspect, in some embodiments of the present invention, the method for using the hash coding algorithm to calculate and obtain the potential stamping areas in the document image to be stamped according to the similarity between the document image to be stamped and the stamping area image template includes the following steps:

[0020] Use hash coding to encode the document image to be stamped and the stamping area image template respectively to obtain the encoded data of the document image to be stamped and the stamping area image template;

[0021] Calculate the similarity between the image of the document to be stamped and the template image of the stamping area using the Euclidean distance calculation formula based on the encoded data of the two, so as to obtain a similarity value;

[0022] Obtain the potential stamping area in the image of the document to be stamped according to the similarity value.

[0023] Based on the first aspect, in some embodiments of the present invention, the method for calculating the entropy value of the filtered potential stamping area includes the following steps:

[0024] Calculate the grayscale value of each pixel point in the filtered potential stamping area;

[0025] Calculate the probability of each grayscale value appearing in the potential stamping area based on the grayscale value of each pixel point to obtain the grayscale probability value;

[0026] Calculate the entropy value of the potential stamping area according to the grayscale probability value.

[0027] Based on the first aspect, in some embodiments of the present invention, the method for calculating the entropy value of the potential stamping area according to the grayscale probability value includes the following steps:

[0028] According to the grayscale probability value, use the formula Calculate the entropy value of the potential stamping area.

[0029] Based on the first aspect, in some embodiments of the present invention, the method for statistically analyzing each comparison result, generating and determining the target stamping area according to the statistical result includes the following steps:

[0030] Statistically analyze and determine whether the number of comparison results exceeding the entropy threshold is less than the preset quantity. If so, determine the potential stamping area as a non-target stamping area; if not, determine the potential stamping area as the target stamping area.

[0031] In the second aspect, an embodiment of the present invention provides a precise stamping area recognition system based on differential dilated convolution verification, including a target selection module, a potential area module, a filtering module, an entropy comparison module, and a target area determination module, where:

[0032] The target selection module is used to select the template image of the stamping area;

[0033] The potential area module is used to input the image of the document to be stamped, calculate using the hash coding algorithm, and obtain the potential stamping area in the image of the document to be stamped according to the similarity between the image of the document to be stamped and the template image of the stamping area;

[0034] A filtering module, which is used to filter the potential areas to be stamped by using convolution kernels with different dilation rates respectively, and calculate the entropy values of the filtered potential areas to be stamped, so as to obtain multiple area entropy values;

[0035] An entropy value comparison module, which is used to compare each area entropy value with a preset entropy value threshold to generate multiple comparison results;

[0036] A target area determination module, which is used to count each comparison result, generate and determine the target stamping area according to the statistical result.

[0037] In order to solve the technical problems of certain errors and low recognition accuracy in the process of identifying the target stamping area in the prior art, this system combines the methods of differential dilated convolution filtering and entropy value verification to accurately identify the stamping area, greatly improving the recognition accuracy for subsequent accurate stamping. First, a large number of stamping area image templates are selected through the target selection module to provide reference data for the follow-up; then, the scanned image of the document to be stamped is input through the potential area module, and the potential areas to be stamped with high similarity to the area image template in the document to be stamped are found by using the hash coding algorithm; then, the filtering module filters the potential areas to be stamped by using convolution kernels with different dilation rates respectively, calculates the entropy value of the area after filtering, and then the entropy value comparison module compares each area entropy value with the preset entropy value threshold to generate multiple comparison results, and the target area determination module counts each comparison result, generates and determines the target stamping area according to the statistical result.

[0038] This system uses the combined method of differential dilated convolution filtering and entropy value verification to improve the accuracy of stamping area recognition. It does not use relatively complex methods or a large amount of training data, reducing the calculation time and resource consumption.

[0039] In a third aspect, an embodiment of the present application provides an electronic device, which includes a memory for storing one or more programs; a processor. When the one or more programs are executed by the processor, the method according to any one of the first aspects above is implemented.

[0040] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method according to any one of the first aspects above is implemented.

[0041] The embodiments of the present invention have at least the following advantages or beneficial effects:

[0042] An embodiment of the present invention provides a method and system for accurately identifying a stamped area based on differential dilated convolution verification, which solves the technical problem of certain errors and low recognition accuracy in the process of identifying the target stamped area in the prior art. The present invention combines the methods of differential dilated convolution filtering and entropy value verification to accurately identify the stamped area, greatly improving the recognition accuracy for subsequent accurate stamping. Without using relatively complex methods or a large amount of training data, the calculation time is reduced and the resource consumption is lowered. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0044] Figure 1 It is a flowchart of a method for accurately identifying a stamped area based on differential dilated convolution verification according to an embodiment of the present invention;

[0045] Figure 2 It is a principle block diagram of a system for accurately identifying a stamped area based on differential dilated convolution verification according to an embodiment of the present invention;

[0046] Figure 3 It is a block diagram of the structure of an electronic device provided by an embodiment of the present invention.

[0047] Icons: 100, target selection module; 200, potential area module; 300, filtering module; 400, entropy value comparison module; 500, target area determination module; 101, memory; 102, processor; 103, communication interface. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0049] Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but merely represents the selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0050] It should be noted that like reference numerals and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it is not necessary to further define and explain it in subsequent figures.

[0051] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.

[0052] In the description of the embodiments of the present invention, "a plurality of" represents at least two.

[0053] Embodiment

[0054] As Figure 1 shown, in a first aspect, an embodiment of the present invention provides a method for accurately identifying a stamped area based on differential dilated convolution verification, including the following steps:

[0055] S1. Select an image template of the stamped area;

[0056] Further, obtain a plurality of initial image template samples of the stamped area, and establish a template database; use a saliency detection method to screen the initial image templates of the stamped area in the template database to obtain an image template of the stamped area.

[0057] Further, use the ST model to perform saliency detection on the initial image templates of the stamped area in the template database to screen and obtain an image template of the stamped area.

[0058] In some embodiments of the present invention, select a sufficient number of representative image templates of the stamped area (such as area images containing important information such as contract signatures, stamping times, company names, etc.); use a saliency detection method to perform secondary screening on the templates, and directly delete the area image templates with poor saliency, leaving only the relatively excellent area image templates. The saliency detection method can use the ST model for saliency detection.

[0059] S2. Input the image of the document to be stamped, and use the hash coding algorithm to calculate and obtain the potential stamping area in the image of the document to be stamped according to the similarity between the image of the document to be stamped and the image template of the stamping area;

[0060] Further, use hash coding to encode the image of the document to be stamped and the image template of the stamping area respectively to obtain the encoded data of the image of the document to be stamped and the image template of the stamping area; calculate the similarity between the image of the document to be stamped and the image template of the stamping area using the Euclidean distance calculation formula based on the encoded data of the image of the document to be stamped and the image template of the stamping area to obtain a similarity value; obtain the potential stamping area in the image of the document to be stamped according to the similarity value.

[0061] In some embodiments of the present invention, input the scanned image of the document to be stamped, and use the hash coding algorithm to find the potential stamping area with a relatively high similarity between the document to be stamped and the area image template. Use hash coding to encode respectively, and use methods such as Euclidean distance to calculate the similarity. The smaller the distance, the higher the similarity.

[0062] S3. Use convolution kernels with different dilation rates to filter the potential stamping area respectively, and calculate the entropy value of the filtered potential stamping area to obtain multiple area entropy values;

[0063] Further, calculate the gray value of each pixel point in the filtered potential stamping area; calculate the probability of each gray level appearing in the potential stamping area based on the gray value of each pixel point to obtain the gray probability value; calculate the entropy value of the potential stamping area according to the gray probability value.

[0064] Further, according to the gray probability value, use the formula to calculate the entropy value of the potential stamping area.

[0065] In some embodiments of the present invention, use a convolution kernel with a dilation rate of 1 to filter the potential stamping area, calculate the entropy value of this area after filtering, denoted as A; use a convolution kernel with a dilation rate of 2 to filter the potential stamping area, calculate the entropy value of this area after filtering, denoted as B; use a convolution kernel with a dilation rate of 3 to filter the potential stamping area, calculate the entropy value of this area after filtering, denoted as C.

[0066] The entropy value calculation method is as follows:

[0067] Calculate the gray value of each pixel point in the area; then calculate the probability of a certain gray level appearing in this image; then use the formula to calculate the area entropy.

[0068] S4. Compare each area entropy value with the preset entropy value threshold to generate multiple comparison results;

[0069] S5. Statistically analyze each comparison result, generate and determine the target stamping area based on the statistical result.

[0070] Furthermore, statistically analyze and determine whether the number of comparison results exceeding the entropy threshold is lower than a preset quantity. If so, determine the potential stamping area as a non-target stamping area; if not, determine the potential stamping area as the target stamping area.

[0071] In some embodiments of the present invention, if no more than 2 of the three entropy values of ABC are lower than the threshold (the pre-set entropy threshold), then the potential stamping area is determined as a non-target stamping area; if more than 2 (including 2) of the three entropy values of ABC are lower than the threshold, the potential stamping area is recognized as the target stamping area, and the stamping device is allowed to stamp in this area.

[0072] To solve the technical problems of certain errors and low recognition accuracy in the process of identifying the target stamping area in the prior art, this method combines the methods of differential dilated convolution filtering and entropy verification to accurately identify the stamping area, greatly improving the recognition accuracy for subsequent accurate stamping. First, select a large number of stamping area image templates to provide reference data for the follow-up; then, input the scanned image of the document to be stamped, and use the hash coding algorithm to find the potential stamping areas in the document to be stamped that have a high similarity with the area image templates; then, use convolution kernels with different dilation rates to filter the potential stamping areas respectively, calculate the entropy value of the area after filtering, and then compare the entropy values of each area with the pre-set entropy threshold to generate multiple comparison results, and statistically analyze each comparison result to generate and determine the target stamping area based on the statistical result.

[0073] This method uses the combined method of differential dilated convolution filtering and entropy verification to improve the accuracy of stamping area recognition. It does not use relatively complex methods or a large amount of training data, reducing the calculation time and resource consumption.

[0074] As Figure 2 shown, in the second aspect, an embodiment of the present invention provides a precise stamping area recognition system based on differential dilated convolution verification, including a target selection module 100, a potential area module 200, a filtering module 300, an entropy comparison module 400, and a target area determination module 500, where:

[0075] The target selection module 100 is used to select stamping area image templates.

[0076] The potential area module 200 is used to input the image of the document to be stamped, and calculate and obtain the potential stamping areas in the image of the document to be stamped according to the similarity between the image of the document to be stamped and the stamping area image templates by using the hash coding algorithm.

[0077] A filtering module 300 is configured to filter potential areas to be stamped using convolutional kernels with different dilation rates respectively, and calculate the entropy values of the filtered potential areas to be stamped, so as to obtain multiple area entropy values;

[0078] An entropy value comparison module 400 is configured to compare each area entropy value with a preset entropy value threshold to generate multiple comparison results;

[0079] A target area determination module 500 is configured to count each comparison result, generate and determine a target stamping area according to the statistical result.

[0080] To solve the technical problems of certain errors and low recognition accuracy in the process of identifying the target stamping area in the prior art, the present system combines the methods of differential atrous convolution filtering and entropy value verification to accurately identify the stamping area, greatly improving the recognition accuracy for subsequent accurate stamping. First, a large number of stamping area image templates are selected through a target selection module 100 to provide reference data for the subsequent process; then, the scanned image of the document to be stamped is input through a potential area module 200, and the potential areas to be stamped with relatively high similarity to the area image templates in the document to be stamped are found using a hash coding algorithm; then, the potential areas to be stamped are filtered respectively by the filtering module 300 using convolutional kernels with different dilation rates, the entropy value of the area is calculated after filtering, and then each area entropy value is compared with a preset entropy value threshold by the entropy value comparison module 400 to generate multiple comparison results, and each comparison result is counted by the target area determination module 500, and the target stamping area is generated and determined according to the statistical result.

[0081] The present system uses a method combining differential atrous convolution filtering and entropy value verification to improve the accuracy of stamping area recognition. It does not use relatively complex methods or a large amount of training data, reducing the calculation time and resource consumption.

[0082] As Figure 3 shown, in a third aspect, an embodiment of the present application provides an electronic device, which includes a memory 101 for storing one or more programs; a processor 102. When the one or more programs are executed by the processor 102, the method according to any one of the first aspects described above is implemented.

[0083] It further includes a communication interface 103. The memory 101, the processor 102, and the communication interface 103 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines. The memory 101 can be used to store software programs and modules. The processor 102 executes various functional applications and data processing by executing the software programs and modules stored in the memory 101. The communication interface 103 can be used for signaling or data communication with other node devices.

[0084] Among them, the memory 101 can be, but is not limited to, a random access memory 101 (Random Access Memory, RAM), a read-only memory 101 (ReadOnly Memory, ROM), a programmable read-only memory 101 (Programmable Read-Only Memory, PROM), an erasable programmable read-only memory 101 (Erasable Programmable Read-Only Memory, EPROM), an electrically erasable programmable read-only memory 101 (Electric Erasable Programmable Read-Only Memory, EEPROM), etc.

[0085] The processor 102 can be an integrated circuit chip with signal processing capabilities. The processor 102 can be a general-purpose processor 102, including a central processing unit 102 (Central Processing Unit, CPU), a network processor 102 (Network Processor, NP), etc.; it can also be a digital signal processor 102 (Digital Signal Processing, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field-programmable gate array (Field-Programmable Gate Array, FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0086] In the embodiments provided in this application, it should be understood that the disclosed methods, systems, and methods can also be implemented in other ways. The method and system embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of the methods, systems, and computer program products according to multiple embodiments of this application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0087] In addition, in each embodiment of this application, the various functional modules can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part.

[0088] Fourthly, the embodiments of this application provide a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by the processor 102, it implements the method according to any one of the above first aspects. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory 101 (ROM, Read-Only Memory), random access memory 101 (RAM, Random Access Memory), magnetic disks, or optical discs that can store program code.

[0089] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

[0090] For those skilled in the art, it is obvious that the present application is not limited to the details of the above-described exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present application. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present application is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application. Any reference signs in the claims should not be construed as limiting the claims involved.

Claims

1. A precise identification method for the sealed area based on differential dilated convolution verification, characterized in that It includes the following steps: Select the image template of the stamping area; It includes obtaining multiple initial stamping area image template samples and establishing a template database; using the saliency detection method to screen the initial stamping area image templates in the template database to obtain the stamping area image template, specifically including using the ST model to perform saliency detection on the initial stamping area image templates in the template database to screen and obtain the stamping area image template; Input the image of the document to be stamped, calculate using the hash coding algorithm and obtain the potential stamping area in the image of the document to be stamped according to the similarity between the image of the document to be stamped and the stamping area image template; Filter the potential stamping area using convolution kernels with different dilation rates respectively, and calculate the entropy value of the filtered potential stamping area to obtain multiple regional entropy values; including calculating the gray value of each pixel point in the filtered potential stamping area; calculating the probability of each gray level appearing in the potential stamping area based on the gray value of each pixel point to obtain the gray probability value; calculating the entropy value of the potential stamping area according to the gray probability value; Compare each regional entropy value with the preset entropy threshold value to generate multiple comparison results; Statistically analyze each comparison result, generate and determine the target stamping area according to the statistical result.

2. The accurate identification method for the stamped area based on differential dilated convolution verification according to claim 1, wherein The method of calculating using the hash coding algorithm and obtaining the potential stamping area in the image of the document to be stamped according to the similarity between the image of the document to be stamped and the stamping area image template includes the following steps: Encode the image of the document to be stamped and the stamping area image template respectively using hash coding to obtain the encoded data of the image of the document to be stamped and the stamping area image template; Calculate the similarity between the image of the document to be stamped and the stamping area image template using the Euclidean distance calculation formula according to the encoded data of the image of the document to be stamped and the stamping area image template to obtain the similarity value; Obtain the potential stamping area in the image of the document to be stamped according to the similarity value.

3. The precise identification method for the sealed area based on differential dilated convolution verification according to claim 1, wherein, The method of calculating the entropy value of the potential stamping area according to the gray probability value includes the following steps: According to the grayscale probability value, use the formula to calculate the entropy value of the potential area to be stamped.

4. A method for accurately identifying a stamped area based on differential dilated convolution verification according to claim 1, characterized in that The method of statistically analyzing each comparison result, generating and determining the target stamping area according to the statistical result includes the following steps: Statistically analyze and judge whether the number of comparison results exceeding the entropy threshold value is lower than the preset quantity. If so, determine the potential stamping area as a non-target stamping area; if not, determine the potential stamping area as the target stamping area.

5. A precise recognition system for the stamped area based on differential dilated convolution verification, characterized in that, It includes a target selection module, a potential area module, a filtering module, an entropy comparison module and a target area determination module, where: The target selection module is used to select the stamping area image template; it includes obtaining multiple initial stamping area image template samples and establishing a template database; using the saliency detection method to screen the initial stamping area image templates in the template database to obtain the stamping area image template, specifically including using the ST model to perform saliency detection on the initial stamping area image templates in the template database to screen and obtain the stamping area image template; A potential area module for inputting an image of a document to be stamped, calculating using a hash coding algorithm, and obtaining potential areas to be stamped in the image of the document to be stamped according to the similarity between the image of the document to be stamped and the image template of the stamping area; A filtering module for filtering the potential areas to be stamped using convolutional kernels with different dilation rates respectively, and calculating the entropy values of the filtered potential areas to be stamped to obtain multiple area entropy values; including calculating the gray value of each pixel point in the filtered potential area to be stamped; calculating the probability of each gray level appearing in the potential area to be stamped based on the gray value of each pixel point to obtain gray probability values; calculating the entropy value of the potential area to be stamped according to the gray probability values; An entropy value comparison module for comparing each area entropy value with a preset entropy value threshold to generate multiple comparison results; A target area determination module for statistically processing each comparison result, generating and determining a target stamping area according to the statistical result.

6. An electronic device, characterized in that, Including: A memory for storing one or more programs; A processor; When the one or more programs are executed by the processor, the method described in any one of claims 1-4 is implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the method described in any one of claims 1-4 is implemented.

Images