OCR recognition effect evaluation method and device

Abstract

The present application discloses an OCR recognition effect evaluation method and device. The OCR recognition effect evaluation method provided by the present application performs optical character recognition on a specific evaluation sample image through an OCR recognition model to be evaluated, obtains model recognition result data output by the OCR recognition model to be evaluated, and compares the model recognition result data with standard recognition result data. When the model recognition result data is closer to the standard recognition result data, the recognition effect of the OCR recognition model is better, thereby obtaining the recognition effect evaluation result of the OCR recognition model, avoiding the influence of subjective factors on the evaluation result, and without going through the stage of user experience, solving the technical problems of inaccurate evaluation results and long evaluation cycle in the existing OCR recognition algorithm evaluation method.

Description

Technical Field

[0001] The present invention relates to the field of optical character recognition technology, and in particular to an OCR recognition effect evaluation method and device. Background Art

[0002] OCR technology (Optical Character Recognition) is currently widely used in related fields such as handwriting recognition, printed text recognition, and text image recognition. From document recognition, license plate recognition, bank card recognition, ID card recognition, bill recognition, invoice recognition, etc. to advertisement and poster content recognition. The invention of OCR technology has greatly simplified the way we process data.

[0003] OCR can be divided into two categories: handwriting recognition and printed text recognition. According to the current OCR recognition capabilities, the accuracy of printed text recognition is much higher than that of handwriting. Although OCR recognition technology is relatively mature and stable, due to the different learning algorithms used by manufacturers, different sets and numbers of trained samples, the recognition effects obtained for the same image or material to be recognized are also different. At present, the evaluation method for OCR recognition effect is mostly the trial experience method, that is, the user actually uses the OCR recognition algorithm and evaluates the recognition effect of the model based on the user experience. This evaluation method has technical problems such as inaccurate evaluation results and long evaluation cycle. Summary of the invention

[0004] The embodiments of the present application provide an OCR recognition effect evaluation method and device for solving the technical problems of inaccurate evaluation results and long evaluation cycle in the existing OCR recognition algorithm evaluation method.

[0005] The first aspect of the present application provides an OCR recognition effect evaluation method, comprising:

[0006] Obtain the OCR recognition model to be evaluated and the evaluation sample image;

[0007] Performing optical character recognition on the evaluation sample image through the OCR recognition model to obtain model recognition result data;

[0008] According to the standard recognition result data corresponding to the evaluation sample image, the model recognition result data is compared with the standard recognition result data to determine the recognition effect evaluation result of the OCR recognition model according to the comparison result.

[0009] Preferably, the process of comparing the model recognition result data with the standard recognition result data specifically includes:

[0010] Comparing the index items in the model recognition result data with the index items in the standard recognition result data to determine the index item recognition comparison result;

[0011] Based on the correspondence between the indicator items and the indicator contents, the indicator contents in the model recognition result data are compared with the indicator contents in the standard recognition result data to determine the indicator content recognition comparison result.

[0012] Preferably, the process of comparing the model recognition result data with the standard recognition result data specifically includes:

[0013] According to the filling mode of the indicator content, the indicator content in the model recognition result data is divided into printed item content and handwritten item content according to the filling type;

[0014] According to two different filling types, the index contents in the model recognition result data are compared with the index contents in the standard recognition result data respectively to obtain the printed item content recognition comparison result and the handwritten item content recognition comparison result.

[0015] Preferably, determining the recognition effect evaluation result of the OCR recognition model according to the comparison result specifically includes:

[0016] The indicator item recognition comparison result and the indicator content recognition comparison result are converted into an indicator item recognition score and an indicator content recognition score respectively, and a recognition effect score is obtained based on the indicator item recognition score and the indicator content recognition score through a summation formula to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score.

[0017] Preferably, determining the recognition effect evaluation result of the OCR recognition model according to the comparison result specifically includes:

[0018] The printed item content recognition comparison result and the handwritten item content recognition comparison result are converted into a printed item content recognition score and a handwritten item content recognition score respectively, and a recognition effect score is obtained based on the printed item content score and the handwritten item content recognition score through a recognition effect score calculation formula, so as to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score, wherein the recognition effect score calculation formula is specifically:

[0019]

[0020] In the formula, O is the recognition effect score, m is the recognition score of the printed item content, n is the recognition score of the handwritten item content, x is the label of the printed item content, and y is the label of the handwritten item content.

[0021] The second aspect of the present application provides an OCR recognition effect evaluation device, comprising:

[0022] An acquisition unit, used for acquiring an OCR recognition model to be evaluated and an evaluation sample image;

[0023] A recognition processing unit, used to perform optical character recognition on the evaluation sample image through the OCR recognition model to obtain model recognition result data;

[0024] The recognition result comparison unit is used to compare the model recognition result data with the standard recognition result data according to the standard recognition result data corresponding to the evaluation sample image, so as to determine the recognition effect evaluation result of the OCR recognition model according to the comparison result.

[0025] Preferably, the recognition result comparison unit is specifically used for:

[0026] Comparing the index items in the model recognition result data with the index items in the standard recognition result data to determine the index item recognition comparison result;

[0027] Based on the correspondence between the indicator items and the indicator contents, the indicator contents in the model recognition result data are compared with the indicator contents in the standard recognition result data to determine the indicator content recognition comparison result.

[0028] Preferably, the recognition result comparison unit is specifically used for:

[0029] According to the filling mode of the indicator content, the indicator content in the model recognition result data is divided into printed item content and handwritten item content according to the filling type;

[0030] According to two different filling types, the index contents in the model recognition result data are compared with the index contents in the standard recognition result data respectively to obtain the printed item content recognition comparison result and the handwritten item content recognition comparison result.

[0031] Preferably, the recognition result comparison unit is specifically used for:

[0032] The indicator item recognition comparison result and the indicator content recognition comparison result are converted into an indicator item recognition score and an indicator content recognition score respectively, and a recognition effect score is obtained based on the indicator item recognition score and the indicator content recognition score through a summation formula to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score.

[0033] Preferably, the recognition result comparison unit is specifically used for:

[0034] The printed item content recognition comparison result and the handwritten item content recognition comparison result are converted into a printed item content recognition score and a handwritten item content recognition score respectively, and a recognition effect score is obtained based on the printed item content score and the handwritten item content recognition score through a recognition effect score calculation formula, so as to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score, wherein the recognition effect score calculation formula is specifically:

[0035]

[0036] In the formula, O is the recognition effect score, m is the recognition score of the printed item content, n is the recognition score of the handwritten item content, x is the label of the printed item content, and y is the label of the handwritten item content.

[0037] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0038] The OCR recognition effect evaluation method provided in the present application performs optical character recognition on a specific evaluation sample image through the OCR recognition model to be evaluated, obtains model recognition result data output by the OCR recognition model to be evaluated, and compares the model recognition result data with the standard recognition result data. The closer the model recognition result data is to the standard recognition result data, the better the recognition effect of the OCR recognition model is, thereby obtaining the recognition effect evaluation result of the OCR recognition model, avoiding the influence of subjective factors on the evaluation result, and without going through the stage of user experience, solving the technical problems of inaccurate evaluation results and long evaluation cycle of existing OCR recognition algorithm evaluation methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0040] Figure 1 A flowchart of an embodiment of an OCR recognition effect evaluation method provided in this application.

[0041] Figure 2 A flowchart of another embodiment of an OCR recognition effect evaluation method provided in the present application.

[0042] Figure 3 A schematic diagram of the structure of an embodiment of an OCR recognition effect evaluation device provided in the present application. DETAILED DESCRIPTION

[0043] The embodiments of the present application provide an OCR recognition effect evaluation method and device for solving the technical problems of inaccurate evaluation results and long evaluation cycle in the existing OCR recognition algorithm evaluation method.

[0044] In order to make the purpose, features, and advantages of the invention of this application more obvious and easy to understand, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the embodiments described below are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0045] See also Figure 1 , this embodiment provides an OCR recognition effect evaluation method, including:

[0046] Step 101: Obtain an OCR recognition model to be evaluated and an evaluation sample image.

[0047] First, obtain the OCR recognition model to be evaluated and the evaluation sample images, where the evaluation sample images can be existing general sample images or sample images created by the user for his own business scenarios, and the OCR recognition model to be evaluated can be an OCR recognition algorithm model designed by the user or an OCR recognition algorithm model purchased from other OCR service providers.

[0048] Step 102: Perform optical character recognition on the evaluation sample image through the OCR recognition model to obtain model recognition result data.

[0049] Next, optical character recognition is performed on the evaluation sample image using the acquired OCR recognition model to obtain model recognition result data output by the OCR recognition model.

[0050] Step 103: According to the standard recognition result data corresponding to the evaluation sample image, the model recognition result data is compared with the standard recognition result data, and the recognition effect evaluation result of the OCR recognition model is determined according to the comparison result.

[0051] Next, based on the standard recognition result data corresponding to the evaluation sample image, where the standard recognition result data can be pre-produced through manual annotation and stored in the database for backup, the model recognition result data is compared with the standard recognition result data. According to the comparison result, the closer the model recognition result data is to the standard recognition result data, the better the recognition effect of the OCR recognition model is, thereby obtaining the recognition effect evaluation result of the OCR recognition model.

[0052] The OCR recognition effect evaluation method provided in this application performs optical character recognition on a specific evaluation sample image through the OCR recognition model to be evaluated, obtains the model recognition result data output by the OCR recognition model to be evaluated, and compares the model recognition result data with the standard recognition result data. When the model recognition result data is closer to the standard recognition result data, the recognition effect of the OCR recognition model is better, thereby obtaining the recognition effect evaluation result of the OCR recognition model, avoiding the influence of subjective factors on the evaluation results, and without going through the stage of user experience, improving resource utilization, and providing a recognition effect quantitative evaluation system for self-built or purchased OCR recognition services from different manufacturers, which helps to reduce investment costs, speed up the progress of technology practical application, and ultimately improve work efficiency. At the same time, it also enhances the stability of OCR service provision. In actual use, multiple OCR recognition services ensure the stability and preferential selectivity of image recognition, and solve the technical problems of inaccurate evaluation results and long evaluation cycles in existing OCR recognition algorithm evaluation methods.

[0053] The above content is a detailed description of an embodiment of an OCR recognition effect evaluation method provided by the present application. The following is a description of a second embodiment of an OCR recognition effect evaluation method provided by the present application.

[0054] Based on the content of the first embodiment above, a second embodiment of the present application provides an OCR recognition effect evaluation method, including:

[0055] For further information, see Figure 2 , the process of comparing the model recognition result data with the standard recognition result data specifically includes:

[0056] Compare the index items in the model recognition result data with the index items in the standard recognition result data to determine the index item recognition comparison result;

[0057] Based on the correspondence between the indicator items and the indicator contents, the indicator contents in the model recognition result data are compared with the indicator contents in the standard recognition result data to determine the indicator content recognition comparison result.

[0058] Figure 2 The bid evaluation index library in can be used to store preset standard identification result data.

[0059] Furthermore, the recognition effect evaluation result of the OCR recognition model is determined according to the comparison result, and specifically includes:

[0060] The indicator item recognition comparison result and the indicator content recognition comparison result are converted into the indicator item recognition score and the indicator content recognition score respectively, and the recognition effect score is obtained through the summation formula based on the indicator item recognition score and the indicator content recognition score, so as to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score.

[0061] It should be noted that this embodiment takes the expansion and registration of the power industry as an example. When applying for the expansion and registration of the power industry, users are often required to provide some application materials. These application materials will be scanned and archived. During the scanning and archiving process, they can be checked for compliance. If unqualified information is found, a message reminder can be issued. In the power marketing business, the image to be recognized is basically a fixed style, but the same material can have different versions and styles. The design of this process converted into a digital model is shown in Table 1:

[0062] Table 1. Example of indicators

[0063]

[0064] The evaluation index content is the record of the identification items on the image to be identified. Taking the sample image of the ID card as an example, as shown in Table 2:

[0065] Table 2. Examples of indicator items of ID cards and corresponding indicator contents

[0066]

[0067] The standard recognition result data of the evaluation sample images of the ID card contains the content corresponding to the index items and index contents shown in the above two tables. For each key index factor, it is divided into two judgment bases:

[0068] The corresponding evaluation result of the indicator item recognition effect is 0 for not recognized and 1 for recognized.

[0069] The indicator item identification corresponds to the recognition effect of the indicator content, and the corresponding evaluation result is 0 for not recognized and 1 for recognized.

[0070] The identification results are assigned values ​​to obtain the indicator item identification score and indicator content identification score. The corresponding assignment expression example is as follows:

[0071] Evaluation results Corresponding value 0 1 1 10

[0072] If the model recognition result data is compared with the standard recognition result data, the closer the indicator items and indicator contents identified by the model recognition result data are to the standard recognition result data, the better the recognition effect is. If the comparison results of the model recognition result data and the standard recognition result data are consistent, it means that the OCR recognition model accurately recognizes the sample image.

[0073] Furthermore, the process of comparing the model recognition result data with the standard recognition result data specifically includes:

[0074] According to the filling method of the indicator content, the indicator content in the model recognition result data is divided into printed item content and handwritten item content according to the filling type;

[0075] According to two different filling types, the index contents in the model recognition result data are compared with the index contents in the standard recognition result data to obtain the printed item content recognition comparison result and the handwritten item content recognition comparison result.

[0076] Furthermore, the recognition effect evaluation result of the OCR recognition model is determined according to the comparison result, and specifically includes:

[0077] The printed item content recognition comparison result and the handwritten item content recognition comparison result are converted into a printed item content recognition score and a handwritten item content recognition score respectively, and the recognition effect score is obtained based on the printed item content score and the handwritten item content recognition score through the recognition effect score calculation formula, so as to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score, wherein the recognition effect score calculation formula is specifically as follows:

[0078]

[0079] Wherein, O is the recognition effect score, m is the recognition score of the printed item content, n is the recognition score of the handwritten item content, x is the label of the printed item content, and y is the label of the handwritten item content.

[0080] It should be noted that this embodiment also provides another way to judge the recognition effect, specifically, according to the indicator content filling method, the indicator content in the model recognition result data is divided into printed item content and handwritten item content according to the filling type; according to the two different filling types, the indicator content in the model recognition result data is compared with the indicator content in the standard recognition result data, and the printed item content recognition comparison result and the handwritten item content recognition comparison result are obtained. After that, the printed item content recognition comparison result and the handwritten item content recognition comparison result can be converted into a printed item content recognition score and a handwritten item content recognition score respectively, and the recognition effect score is obtained based on the printed item content score and the handwritten item content recognition score through the recognition effect score calculation formula, so as to realize the recognition effect evaluation of the OCR recognition model for different fonts.

[0081] The above content is a detailed description of a second embodiment of an OCR recognition effect evaluation method provided by the present application. The following is a detailed description of an embodiment of an OCR recognition effect evaluation device provided by the present application.

[0082] See also Figure 3 The third embodiment of the present application provides an OCR recognition effect evaluation device, comprising:

[0083] An acquisition unit 201 is used to acquire an OCR recognition model to be evaluated and an evaluation sample image;

[0084] The recognition processing unit 202 is used to perform optical character recognition on the evaluation sample image through the OCR recognition model to obtain model recognition result data;

[0085] The recognition result comparison unit 203 is used to compare the model recognition result data with the standard recognition result data according to the standard recognition result data corresponding to the evaluation sample image, so as to determine the recognition effect evaluation result of the OCR recognition model according to the comparison result.

[0086] Furthermore, the recognition result comparison unit is specifically used for:

[0087] Compare the index items in the model recognition result data with the index items in the standard recognition result data to determine the index item recognition comparison result;

[0088] Based on the correspondence between the indicator items and the indicator contents, the indicator contents in the model recognition result data are compared with the indicator contents in the standard recognition result data to determine the indicator content recognition comparison result.

[0089] Furthermore, the recognition result comparison unit is specifically used for:

[0090] According to the filling method of the indicator content, the indicator content in the model recognition result data is divided into printed item content and handwritten item content according to the filling type;

[0091] According to two different filling types, the index contents in the model recognition result data are compared with the index contents in the standard recognition result data to obtain the printed item content recognition comparison result and the handwritten item content recognition comparison result.

[0092] Furthermore, the recognition result comparison unit is specifically used for:

[0093] The indicator item recognition comparison result and the indicator content recognition comparison result are converted into the indicator item recognition score and the indicator content recognition score respectively, and the recognition effect score is obtained through the summation formula based on the indicator item recognition score and the indicator content recognition score, so as to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score.

[0094] Furthermore, the recognition result comparison unit is specifically used for:

[0095] The printed item content recognition comparison result and the handwritten item content recognition comparison result are converted into a printed item content recognition score and a handwritten item content recognition score respectively, and the recognition effect score is obtained based on the printed item content score and the handwritten item content recognition score through the recognition effect score calculation formula, so as to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score, wherein the recognition effect score calculation formula is specifically as follows:

[0096]

[0097] Wherein, O is the recognition effect score, m is the recognition score of the printed item content, n is the recognition score of the handwritten item content, x is the label of the printed item content, and y is the label of the handwritten item content.

[0098] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the terminals, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0099] In the several embodiments provided in the present application, it should be understood that the disclosed terminals, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0100] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can, for example, be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0101] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0102] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0103] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0104] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for evaluating OCR recognition effect, characterized in that: include: Obtain the OCR recognition model to be evaluated and the evaluation sample image; Performing optical character recognition on the evaluation sample image through the OCR recognition model to obtain model recognition result data; According to the standard recognition result data corresponding to the evaluation sample image, the model recognition result data is compared with the standard recognition result data to determine the recognition effect evaluation result of the OCR recognition model according to the comparison result; The process of comparing the model recognition result data with the standard recognition result data specifically includes: According to the filling mode of the indicator content, the indicator content in the model recognition result data is divided into printed item content and handwritten item content according to the filling type; According to two different filling types, the index contents in the model recognition result data are compared with the index contents in the standard recognition result data to obtain a printed item content recognition comparison result and a handwritten item content recognition comparison result; Determining the recognition effect evaluation result of the OCR recognition model according to the comparison result specifically includes: The printed item content recognition comparison result and the handwritten item content recognition comparison result are converted into a printed item content recognition score and a handwritten item content recognition score respectively, and a recognition effect score is obtained based on the printed item content score and the handwritten item content recognition score through a recognition effect score calculation formula, so as to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score, wherein the recognition effect score calculation formula is specifically: ； In the formula, O is the recognition effect score, m is the recognition score of the printed item content, n is the recognition score of the handwritten item content, x is the label of the printed item content, and y is the label of the handwritten item content.

2. The OCR recognition effect evaluation method according to claim 1, characterized in that: The process of comparing the model recognition result data with the standard recognition result data specifically includes: Comparing the index items in the model recognition result data with the index items in the standard recognition result data to determine the index item recognition comparison result; Based on the correspondence between the indicator items and the indicator contents, the indicator contents in the model recognition result data are compared with the indicator contents in the standard recognition result data to determine the indicator content recognition comparison result.

3. The OCR recognition effect evaluation method according to claim 2, characterized in that: Determining the recognition effect evaluation result of the OCR recognition model according to the comparison result specifically includes: The indicator item recognition comparison result and the indicator content recognition comparison result are converted into an indicator item recognition score and an indicator content recognition score respectively, and a recognition effect score is obtained based on the indicator item recognition score and the indicator content recognition score through a summation formula to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score.

4. An OCR recognition effect evaluation device, characterized in that: include: An acquisition unit, used for acquiring an OCR recognition model to be evaluated and an evaluation sample image; A recognition processing unit, used to perform optical character recognition on the evaluation sample image through the OCR recognition model to obtain model recognition result data; A recognition result comparison unit, configured to compare the model recognition result data with the standard recognition result data according to the standard recognition result data corresponding to the evaluation sample image, so as to determine the recognition effect evaluation result of the OCR recognition model according to the comparison result; The recognition result comparison unit is specifically used for: According to the filling mode of the indicator content, the indicator content in the model recognition result data is divided into printed item content and handwritten item content according to the filling type; According to two different filling types, the index contents in the model recognition result data are compared with the index contents in the standard recognition result data to obtain a printed item content recognition comparison result and a handwritten item content recognition comparison result; The recognition result comparison unit is specifically used for: The printed item content recognition comparison result and the handwritten item content recognition comparison result are converted into a printed item content recognition score and a handwritten item content recognition score respectively, and a recognition effect score is obtained based on the printed item content score and the handwritten item content recognition score through a recognition effect score calculation formula, so as to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score, wherein the recognition effect score calculation formula is specifically: ； In the formula, O is the recognition effect score, m is the recognition score of the printed item content, n is the recognition score of the handwritten item content, x is the label of the printed item content, and y is the label of the handwritten item content.

5. The OCR recognition effect evaluation device according to claim 4, characterized in that: The recognition result comparison unit is specifically used for: Comparing the index items in the model recognition result data with the index items in the standard recognition result data to determine the index item recognition comparison result; Based on the correspondence between the indicator items and the indicator contents, the indicator contents in the model recognition result data are compared with the indicator contents in the standard recognition result data to determine the indicator content recognition comparison result.

6. The OCR recognition effect evaluation device according to claim 5, characterized in that: The recognition result comparison unit is specifically used for: The indicator item recognition comparison result and the indicator content recognition comparison result are converted into an indicator item recognition score and an indicator content recognition score respectively, and a recognition effect score is obtained based on the indicator item recognition score and the indicator content recognition score through a summation formula to determine the recognition effect evaluation result of the OCR recognition model according to the recognition effect score.

Images