Image recognition method and device for paper marking system, computer equipment and storage medium

Abstract

The invention provides an image recognition method and device for a paper marking system, computer equipment and a storage medium. The method comprises the steps of obtaining a candidate character set and reference answer information of a target subject; the candidate character set comprises all candidate characters; deleting part of candidate characters in the candidate character set according to the reference answer information to obtain a target character set; adjusting an initial recognition model according to the target character set to generate a first target recognition model; and identifying a to-be-identified answer image through the first target identification model to obtain a first identification result. According to the embodiment of the invention, the character recognition range can be reduced, so that the image recognition accuracy and recognition efficiency of examination paper marking are remarkably improved, and the paper marking efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of image recognition technology, and in particular to an image recognition method, apparatus, computer equipment, and storage medium for a paper marking system. Background Art

[0002] With the rapid development of educational informatization and artificial intelligence technology, intelligent marking systems have gradually replaced traditional manual marking methods and are widely used in various examinations at all levels, significantly improving marking efficiency.

[0003] Intelligent grading systems use optical character recognition (OCR) technology to automatically identify and grade handwritten answers on exam answer sheets or test papers. However, OCR systems typically need to identify and compare complete sets of Chinese characters. Due to the large number of Chinese characters, the recognition range is too large. Furthermore, in cases of irregular handwriting, blurred handwriting, or significant font variations, the recognition error rate increases significantly, reducing the accuracy and credibility of the recognition results and seriously affecting the objectivity and accuracy of the grading process. Summary of the Invention

[0004] In view of this, the present application proposes an image recognition method, apparatus, computer equipment and storage medium for a marking system to solve the problem in related technologies of increased recognition error rate due to excessive recognition range.

[0005] The first embodiment of the present application provides an image recognition method for a paper marking system, the method comprising:

[0006] Obtaining a candidate character set and reference answer information for a target subject; the candidate character set includes all candidate characters;

[0007] deleting some candidate characters in the candidate character set according to the reference answer information to obtain a target character set;

[0008] Adjusting the initial recognition model according to the target character set to generate a first target recognition model;

[0009] The answer image to be identified is identified by using the first target recognition model to obtain a first recognition result.

[0010] In an embodiment of the present application, the reference answer information is a first character set corresponding to the target subject; the first character set includes multiple first characters; the first characters are words, symbols, or numbers; and the target character set is obtained by deleting some candidate characters from the candidate character set using the reference answer information, including:

[0011] For any first character in the first character set, calculating multiple matching degrees between the first character and multiple candidate characters in the candidate character set;

[0012] Filtering at least one target character corresponding to the first character from the plurality of candidate characters according to the plurality of matching degrees; the target character refers to a candidate character with a matching degree higher than a preset threshold among the plurality of candidate characters;

[0013] Multiple second character sets are generated based on multiple target characters corresponding to multiple first characters; each second character set includes at least one target character corresponding to the corresponding first character; the target character set is one or more of the multiple second character sets.

[0014] The embodiment of the present application filters out at least one target character corresponding to the first character from the multiple candidate characters through multiple matching degrees; the target character refers to the candidate character with a matching degree higher than a preset threshold among the multiple candidate characters; and generates multiple second character sets based on the multiple target characters corresponding to the multiple first characters, which can greatly reduce the character recognition range of the first target recognition model, thereby reducing the recognition error rate, and thus significantly improving the reliability and efficiency of the intelligent marking system in the actual examination environment, reducing the workload of manual review, and improving the overall marking speed.

[0015] In an embodiment of the present application, adjusting the initial recognition model according to the target character set to generate a first target recognition model includes:

[0016] The target character set is adjusted according to the respondent information to obtain a third character set; the respondent information includes Chinese character mastery level information and respondent experience information; the number of characters in the third character set is less than the number of characters in the target character set;

[0017] The initial recognition model is optimized according to the third character set to generate the first target recognition model.

[0018] In an embodiment of the present application, the first recognition result includes a plurality of fourth characters; after obtaining the first recognition result, the method further includes:

[0019] For any fourth character among the plurality of fourth characters, calculating the confidence of the fourth character;

[0020] Determining whether the confidence level is less than a confidence threshold;

[0021] If the confidence level is less than a confidence threshold, marking the fourth character as a misrecognized character;

[0022] A feedback report is generated according to the misrecognized characters and context information of the misrecognized characters, and the feedback report is sent to a marking end.

[0023] In an embodiment of the present application, calculating multiple matching degrees between the first character and multiple candidate characters in the candidate character set includes:

[0024] Calculate the similarity between the semantic feature vector of the first character and the semantic feature vectors of multiple candidate characters; and use the similarity as the matching degree.

[0025] In an embodiment of the present application, before the first object recognition model is used to recognize the answer image to be recognized, the method further includes:

[0026] Get the initial answer image;

[0027] Performing denoising processing on the initial answer image to obtain a first image; the denoising processing is to replace each pixel value in the initial answer image with the median value in its neighborhood;

[0028] Performing handwriting correction processing on the first image to obtain a second image; the handwriting correction processing refers to removing broken or blurred parts of character strokes in the first image;

[0029] The character outline clearing process is performed on the second image to obtain a third image; the character outline clearing process refers to enhancing edge information and outline information of the second image.

[0030] In an embodiment of the present application, the method further includes:

[0031] For any first character, calculating the word frequency of the first character in the reference answer information and extracting a semantic feature vector of the first character;

[0032] determining a recognition weight of the first character based on the word frequency and context information of the first character; the context information is determined based on the semantic feature vector of the first character and the semantic feature vectors of adjacent first characters;

[0033] adjusting the initial recognition model according to a plurality of recognition weights of a plurality of first characters in the first character set to generate a second target recognition model;

[0034] Recognize the answer image to be recognized by the second object recognition model to obtain a second recognition result;

[0035] A final recognition result is determined according to the first recognition result and the second recognition result.

[0036] The embodiment of the present application determines the recognition weight of the first character through the word frequency and context information of the first character, and adjusts the initial recognition model according to the multiple recognition weights of multiple first characters in the first character set to generate a second target recognition model. This can significantly reduce the recognition range of the second target recognition model, thereby reducing the recognition error rate, and significantly improving the reliability and efficiency of the intelligent marking system in the actual examination environment, reducing the workload of manual review, and improving the overall marking speed.

[0037] An embodiment of the second aspect of the present application provides an image recognition device for a paper marking system, comprising:

[0038] An information acquisition module, configured to acquire a candidate character set and reference answer information for a target subject; the candidate character set includes all candidate characters;

[0039] a target character set determination module, configured to delete some candidate characters from the candidate character set according to the reference answer information to obtain a target character set;

[0040] A model adjustment module, configured to adjust the initial recognition model according to the target character set to generate a first target recognition model;

[0041] The recognition module is used to recognize the answer image to be recognized through the first target recognition model to obtain a first recognition result.

[0042] An embodiment of the third aspect of the present application provides a computer device, which includes a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the image recognition method for the paper marking system described in the first aspect by executing the computer instructions.

[0043] An embodiment of the fourth aspect of the present application provides a computer-readable storage medium having computer instructions stored thereon, wherein the computer instructions are used to enable a computer to execute the image recognition method for a paper marking system described in the first aspect above.

[0044] Additional aspects and advantages of the present application will be given in part in the description below and in part will become apparent from the description below or learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present application. Throughout the accompanying drawings, the same reference numerals are used to denote the same components.

[0046] In the attached figure:

[0047] Figure 1 A schematic flow chart of an image recognition method for a paper marking system provided in one embodiment of the present application is shown;

[0048] Figure 2 A schematic structural diagram of an image recognition device for a paper marking system provided by an embodiment of the present application is shown;

[0049] Figure 3 A schematic diagram of the structure of a computer device provided in one embodiment of the present application is shown;

[0050] Figure 4 A schematic diagram of a storage medium provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0051] The following describes exemplary embodiments of the present application in more detail with reference to the accompanying drawings. Although exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. Instead, these embodiments are provided to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.

[0052] It should be noted that, unless otherwise specified, the technical or scientific terms used in this application should have the common meanings understood by those skilled in the art to which this application belongs.

[0053] The following describes the technical scenarios involved in the embodiments of this application.

[0054] Most existing OCR technologies use a universal character recognition dictionary for recognition. This universal dictionary covers all commonly used Chinese characters and symbols, and therefore cannot accurately match specific exam questions. Especially in applications in the field of education, exam questions are usually fixed and clear, and the Chinese characters and symbols that appear in the questions generally have strong regularity. In response to the shortcomings of the existing technology, the present invention proposes a technical solution that can automatically narrow the OCR recognition dictionary range and optimize character weights based on reference answers or the range of Chinese characters mastered by students, thereby improving the accuracy of handwriting recognition in intelligent marking systems.

[0055] According to an embodiment of the present application, an embodiment of an image recognition method for a paper marking system is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0056] In this embodiment, an image recognition method for a marking system is provided, which can be used in an intelligent marking system. Figure 1 is a flow chart of an image recognition method for a paper marking system according to an embodiment of the present application. Figure 1 As shown, the process includes the following steps:

[0057] Step S101: Acquire a candidate character set and reference answer information of a target subject.

[0058] Among them, the candidate character set includes all candidate characters; the reference answer information is the first character set corresponding to the target subject; the first character set includes multiple first characters; the first characters are words, symbols or numbers.

[0059] In the embodiment of the present application, the target subjects include but are not limited to Chinese, mathematics, English, science-related subjects (physics, chemistry, biology), and liberal arts-related subjects (politics, history, and geography). The standard reference answers for the corresponding examination subjects can be obtained. Among them, the multiple first characters included in the first character set can be understood as the answers to multiple test questions in a test paper. For example, if the answer to a fill-in-the-blank question is "Hai Nian Wu Zhi Ji", then "Hai Nian Wu Zhi Ji" can be used as a first character.

[0060] Step S102: deleting some candidate characters in the candidate character set according to the reference answer information to obtain a target character set.

[0061] In some specific embodiments, the above step S102 includes steps S1021 to S1023:

[0062] Step S1021 : For any first character in the first character set, calculate multiple matching degrees between the first character and multiple candidate characters in the candidate character set.

[0063] In an embodiment of the present application, the degree of matching between the first character and the candidate characters can be calculated by first determining the semantic feature vector of the first character and the semantic feature vector of each candidate character, and then using cosine similarity or Euclidean distance to calculate the degree of matching between the semantic feature vector of each candidate character and the semantic feature vector of the first character, that is:

[0064]

[0065] Among them, x i represents the semantic feature vector of the candidate character, and y represents the semantic feature vector of the first character.

[0066] Step S1022: Filter out at least one target character corresponding to the first character from the multiple candidate characters according to the multiple matching degrees.

[0067] In the embodiment of the present application, the target character refers to a candidate character among the plurality of candidate characters whose matching degree is higher than a preset threshold value, wherein the preset threshold value can be set according to the situation and is not specifically limited here.

[0068] In the embodiment of the present application, the N characters with the highest scores can be selected from the candidate set according to a set threshold or by using a Top-K screening mechanism.

[0069] Step S1023: Generate multiple second character sets according to the multiple target characters corresponding to the multiple first characters.

[0070] In an embodiment of the present application, each second character set includes at least one target character corresponding to the corresponding first character; the target character set is one or more of the multiple second character sets.

[0071] Step S103: adjusting the initial recognition model according to the target character set to generate a first target recognition model.

[0072] In this embodiment of the present application, the target character set can be adjusted based on respondent information to obtain a third character set. The respondent information includes information about their Chinese character proficiency and their experience. The third character set has fewer characters than the target character set. The initial recognition model is then optimized based on the third character set to generate the first target recognition model.

[0073] In the embodiment of the present application, the respondent information may also include the student's region and the student's school.

[0074] The embodiment of the present application further filters the second character in the target character set through the respondent information, and can eliminate characters that are not related to the respondent information, thereby improving the OCR recognition accuracy of the test paper marking.

[0075] Step S104: Recognize the answer image to be recognized by using the first object recognition model to obtain a first recognition result.

[0076] Before step S104, the method further includes steps a1 to a4:

[0077] Step a1: Obtain the initial answer image.

[0078] Step a2: performing denoising processing on the initial answer image to obtain a first image.

[0079] In the embodiment of the present application, the denoising process refers to replacing each pixel value in the initial answer image with the median value in its neighborhood. Other denoising methods may also be adopted, which are not specifically limited here.

[0080] Step a3: Perform handwriting correction processing on the first image to obtain a second image.

[0081] In an embodiment of the present application, the handwriting correction processing refers to removing the broken parts or blurred parts of the character strokes in the first image.

[0082] Step a4: performing character outline clearing processing on the second image to obtain a third image.

[0083] In the embodiment of the present application, the character outline clearing processing refers to enhancing the edge information and outline information of the second image.

[0084] The embodiments of the present application can effectively improve the basic recognition capabilities of OCR by applying multiple image enhancement algorithms, including character outline sharpening, handwriting correction and denoising technology.

[0085] In some specific embodiments, the first recognition result includes a plurality of fourth characters; after obtaining the first recognition result, the method further includes steps b1 to b4:

[0086] Step b1: calculating the confidence of any fourth character among the plurality of fourth characters.

[0087] In the embodiments of this application, calculating the confidence level for each character is an important step in assessing the reliability of the recognition results. Confidence is typically a value between 0 and 1 that indicates the model's confidence in the recognition result. In deep learning-based OCR systems, the model typically outputs a predicted probability for each character. These probabilities can be directly used as the character's confidence level.

[0088] Step b2: Determine whether the confidence level is less than a confidence threshold.

[0089] In the embodiment of the present application, the confidence threshold can be set according to actual conditions and is not specifically limited here.

[0090] Step b3: if the confidence level is less than the confidence threshold, marking the fourth character as a misrecognized character;

[0091] Step b4: generating a feedback report based on the misrecognized characters and context information of the misrecognized characters, and sending the feedback report to the examination paper grading end.

[0092] The embodiment of the present application performs confidence analysis on each recognized character, automatically marks low-confidence characters as suspected misrecognitions, and generates a feedback report for manual examiners to quickly verify, thereby improving the accuracy of examination.

[0093] In some specific embodiments, the method further includes steps 201 to 205:

[0094] Step 201 : for any first character, calculate the word frequency of the first character in the reference answer information and extract the semantic feature vector of the first character.

[0095] In the embodiment of the present application, each first character in the first character set is a word, a symbol or a number. For example, the reference answer information of a math test is "Solution: x 2 +y 2 =25", the first character set extracted is [solution,:,x, 2 ,+,y, 2 , =, 2, 5]. The extracted first character set can be used as a special OCR recognition dictionary for the target subject test.

[0096] In the embodiment of the present application, word frequency refers to the frequency of occurrence, which can be understood as the frequency of occurrence of the first character in the reference answer information. The semantic feature vector is used to represent the semantic information of the first character.

[0097] In some specific embodiments, the word frequency of the first character can be calculated by traversing the reference answer information, and for each first character, checking whether the first character is already in a counter; if so, incrementing the count of the first character by one; if not, adding the first character to the counter and setting the count to 1.

[0098] In some specific embodiments, the semantic feature vector of the first character may be obtained by:

[0099] Using semantic feature extraction methods in NLP technology (such as word embedding or sentence embedding), each character in the reference answer information is mapped to a high-dimensional vector space, thereby obtaining the semantic feature vector corresponding to each character. These vectors can capture the semantic information of the character.

[0100] In some specific embodiments, the specific method for obtaining the semantic feature vector of the first character is not specifically limited, and the semantic feature vector of the first character can be obtained. For example, the semantic feature vector corresponding to each character can be calculated by combining a transformer model with a model commonly used in traditional NLP.

[0101] Before step S202, the method further includes: determining context information of the first character based on the semantic feature vector of the first character and the semantic feature vectors of adjacent first characters, and the specific determination method is as follows:

[0102] Calculate the similarity between the semantic feature vector of the first character and the semantic feature vector of an adjacent first character; and use the similarity as context information of the first character.

[0103] In an embodiment of the present application, the similarity calculation can adopt cosine similarity, which measures the similarity of two vectors in direction. The closer the value is to 1, the more similar they are. The calculated similarity is used as the context information of the first character to adjust the recognition weight of each first character in OCR recognition. If the similarity of two first characters is high, it means that they may be more semantically related, which can serve as context information to help the OCR system better understand the text.

[0104] The above embodiment is described with examples:

[0105] Assume that there is "x 2 +y 2 =z 2 ”;

[0106] Extract the semantic feature vector of character x and calculate the semantic feature vector of character x and the adjacent characters " 2 " similarity of the semantic feature vector; when the similarity value is greater than the similarity threshold, it means that "x" is often followed by " 2 ", therefore, the recognition weight of character x can be improved.

[0107] Specific recognition scenario: When the OCR system recognizes an ambiguous character, if it is likely to be "x" and it is followed by a superscript symbol "^", then the recognition weight of "x" is increased.

[0108] If the OCR system recognizes an ambiguous character that may be "y" or "z", and it is preceded by a "+" sign, then the recognition weight of "y" or "z" is also increased.

[0109] Step 202: Determine the recognition weight of the first character according to the word frequency and context information of the first character.

[0110] The context information is determined based on the semantic feature vector of the first character and the semantic feature vector of the adjacent first character. The adjacent first character can be understood as the character on the left or the character on the right of the first character.

[0111] In some specific embodiments, the recognition weight of the first character may be calculated as follows:

[0112] The recognition weight of the first character is calculated based on the word frequency, the first weight coefficient corresponding to the word frequency, the context information, and the second weight coefficient corresponding to the context information, as shown in the following formula:

[0113] a1*A+b1*B=C

[0114] Among them, A represents the word frequency of the first character, a1 represents the first weight coefficient corresponding to the word frequency A, B represents the context information of the first character, b1 represents the second weight coefficient corresponding to the context information B, and C represents the recognition weight of the first character. The above-mentioned first weight coefficient and second weight coefficient can be set according to actual conditions and are not specifically limited here.

[0115] Step 203: Adjust the initial recognition model according to the multiple recognition weights of the multiple first characters in the first character set to generate a second target recognition model.

[0116] In an embodiment of the present application, in an OCR (optical character recognition) system, by adjusting the recognition weights of characters in the OCR recognition dictionary (i.e., the first character set mentioned above), the recognition performance of the model can be optimized to better adapt to specific scenarios or needs.

[0117] In an embodiment of the present application, the initial OCR model is fine-tuned using a first set of characters with weights. During the fine-tuning process, the initial OCR model adjusts its recognition tendency for different characters based on the weights (for example, by modifying hyperparameters in the configuration file, such as the learning rate and batch size, to accommodate the fine-tuning process), thereby improving the recognition accuracy of the target characters.

[0118] Step 204: Recognize the answer image to be recognized by the second object recognition model to obtain a second recognition result.

[0119] In the embodiment of the present application, the first object recognition model and the second object recognition model include but are not limited to Tesseract, EasyOCR, and PaddleOCR. The first object recognition model and the second object recognition model are any one of the above multiple recognition models, and the first object recognition model and the second object recognition model are different.

[0120] Step 205: Determine a final recognition result based on the first recognition result and the second recognition result.

[0121] In the embodiment of the present application, the first recognition result and the second recognition module are preliminarily integrated, and a voting mechanism or an ensemble learning algorithm, such as weighted average or rule-based fusion, can be used.

[0122] The embodiment of the present application uses hybrid model integration technology to combine the recognition advantages of multiple recognition sub-models, which can significantly improve the accuracy and robustness of OCR recognition. This method is particularly suitable for processing complex scenes, multilingual environments or low-quality images.

[0123] In some specific embodiments, in an actual system, in addition to the first character set and the second character set, if the confidence of the top 1 / top 5 characters recognized in the first / second character set is not high, this situation is relatively rare. The embodiment of the present application will use full character set processing to ensure the accuracy of recognition. Although in this case, the accuracy of marking is generally not affected, because the students' answers are likely to be wrong, it is only to improve the OCR recognition rate and improve the user experience.

[0124] Corresponding to the implementation of the above image recognition method for the paper marking system, the embodiment of the present application further provides an image recognition device for the paper marking system, which is used to execute the image recognition method for the paper marking system described in the above embodiment. Figure 2 As shown, the image recognition device for the examination paper marking system includes:

[0125] An information acquisition module, configured to acquire a candidate character set and reference answer information for a target subject; the candidate character set includes all candidate characters;

[0126] a target character set determination module, configured to delete some candidate characters from the candidate character set according to the reference answer information to obtain a target character set;

[0127] A model adjustment module, configured to adjust the initial recognition model according to the target character set to generate a first target recognition model;

[0128] The recognition module is used to recognize the answer image to be recognized through the first target recognition model to obtain a first recognition result.

[0129] Optionally, the reference answer information is a first character set corresponding to the target subject; the first character set includes multiple first characters; the first character is a word, a symbol or a number; the target character set determination module is also used to calculate, for any first character in the first character set, multiple matching degrees between the first character and multiple candidate characters in the candidate character set; at least one target character corresponding to the first character is filtered out from the multiple candidate characters based on the multiple matching degrees; the target character refers to a candidate character in the multiple candidate characters whose matching degree is higher than a preset threshold; multiple second character sets are generated based on multiple target characters corresponding to multiple first characters; each second character set includes at least one target character corresponding to the corresponding first character; the target character set is one or more of the multiple second character sets.

[0130] Optionally, the model adjustment module is further used to adjust the target character set according to the respondent information to obtain a third character set; the respondent information includes Chinese character mastery level information and respondent experience information; the number of characters in the third character set is less than the number of characters in the target character set; and the initial recognition model is optimized according to the third character set to generate the first target recognition model.

[0131] Optionally, the first recognition result includes a plurality of fourth characters; and the device further includes:

[0132] The confidence judgment module is used to calculate the confidence of any one of the multiple fourth characters after obtaining the first recognition result; judge whether the confidence is less than a confidence threshold; if the confidence is less than the confidence threshold, mark the fourth character as an incorrectly recognized character; generate a feedback report based on the incorrectly recognized character and the context information of the incorrectly recognized character, and send the feedback report to the marking end.

[0133] Optionally, the device also includes: an image processing module, used to obtain an initial answer image before the answer image to be identified is identified by the first target recognition model; performing denoising processing on the initial answer image to obtain a first image; the denoising processing refers to replacing each pixel value in the initial answer image with the median value within its neighborhood; performing handwriting correction processing on the first image to obtain a second image; the handwriting correction processing refers to removing broken parts or blurred parts of the character strokes in the first image; performing character contour clearing processing on the second image to obtain a third image; the character contour clearing processing refers to enhancing the edge information and contour information of the second image.

[0134] Optionally, the device also includes: a final recognition result determination module, which is used to calculate the word frequency of any first character in the reference answer information and extract the semantic feature vector of the first character; determine the recognition weight of the first character based on the word frequency of the first character and context information; the context information is determined based on the semantic feature vector of the first character and the semantic feature vector of the adjacent first character; adjust the initial recognition model based on multiple recognition weights of multiple first characters in the first character set to generate a second target recognition model; recognize the answer image to be recognized through the second target recognition model to obtain a second recognition result; determine the final recognition result based on the first recognition result and the second recognition result.

[0135] The image recognition device for the paper marking system provided in the above-mentioned embodiment of the present application and the image recognition method for the paper marking system provided in the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0136] The present application also provides a computer device to execute the above-mentioned image recognition method for the examination paper marking system. Figure 3 , which shows a schematic diagram of a computer device provided by some embodiments of the present application. Figure 3 As shown, the computer device 3 includes: a processor 300, a memory 301, a bus 302 and a communication interface 303. The processor 300, the communication interface 303 and the memory 301 are connected via the bus 302; the memory 301 stores a computer program that can be run on the processor 300, and when the processor 300 runs the computer program, it executes the image recognition method for the paper marking system provided in the aforementioned embodiment of the present application.

[0137] The memory 301 may include high-speed random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage. The system network element and at least one other network element are connected via at least one communication interface 303 (which may be wired or wireless), and may use the Internet, a wide area network, a local area network, a metropolitan area network, or the like.

[0138] Bus 302 may be an ISA bus, a PCI bus, or an EISA bus. Such buses may be classified as address buses, data buses, and control buses. Memory 301 is used to store programs, and processor 300 executes the programs upon receiving execution instructions. The image recognition method for a paper marking system disclosed in the aforementioned embodiments may be applied to or implemented by processor 300.

[0139] The processor 300 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by hardware integrated logic circuits in the processor 300 or by software instructions. The above processor 300 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in conjunction with the embodiments of this application can be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium mature in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in the memory 301 , and the processor 300 reads the information in the memory 301 and completes the steps of the above method in combination with its hardware.

[0140] The computer device provided in the embodiment of the present application and the image recognition method for the paper marking system provided in the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, operated or implemented by them.

[0141] The present application also provides a computer-readable storage medium corresponding to the image recognition method for the examination paper marking system provided in the above embodiment. Figure 4 The computer-readable storage medium shown is a CD 30 on which a computer program (i.e., a program product) is stored. When the computer program is run by the processor, it will execute the image recognition method for the paper marking system provided by any of the aforementioned embodiments.

[0142] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical or magnetic storage media, which are not listed here one by one.

[0143] The computer-readable storage medium provided in the above-mentioned embodiments of the present application and the image recognition method for the examination paper system provided in the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0144] It should be noted that:

[0145] In the description provided herein, a large number of specific details are described. However, it is understood that the embodiments of the present application can be practiced without these specific details. In some instances, well-known structures and technologies are not shown in detail so as not to obscure the understanding of this description.

[0146] Similarly, it should be understood that in order to streamline the present application and aid in understanding one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present application, various features of the present application are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the following schematic diagram: the claimed application requires more features than the features expressly recited in each claim. Rather, as reflected in the claims below, inventive aspects lie in less than all the features of the individual embodiments disclosed above. Therefore, the claims following the detailed description are hereby expressly incorporated into the detailed description, with each claim itself serving as a separate embodiment of the present application.

[0147] Furthermore, those skilled in the art will appreciate that although some embodiments described herein include certain features included in other embodiments but not other features, combinations of features from different embodiments are intended to be within the scope of this application and to form different embodiments. For example, in the claims below, any of the claimed embodiments may be used in any combination.

[0148] The above description is merely a preferred embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. An image recognition method for a paper marking system, characterized in that: The method comprises: Obtaining a candidate character set and reference answer information for a target subject; the candidate character set includes all candidate characters; deleting some candidate characters in the candidate character set according to the reference answer information to obtain a target character set; Adjusting the initial recognition model according to the target character set to generate a first target recognition model; The answer image to be identified is identified by using the first target recognition model to obtain a first recognition result.

2. The method according to claim 1, characterized in that The reference answer information is a first character set corresponding to the target subject; the first character set includes a plurality of first characters; the first characters are words, symbols or numbers; Deleting some candidate characters from the candidate character set according to the reference answer information to obtain a target character set includes: For any first character in the first character set, calculating multiple matching degrees between the first character and multiple candidate characters in the candidate character set; Filtering at least one target character corresponding to the first character from the plurality of candidate characters according to the plurality of matching degrees; the target character refers to a candidate character with a matching degree higher than a preset threshold among the plurality of candidate characters; Multiple second character sets are generated based on multiple target characters corresponding to multiple first characters; each second character set includes at least one target character corresponding to the corresponding first character; the target character set is one or more of the multiple second character sets.

3. The method according to claim 1 or 2, characterized in that Adjusting the initial recognition model according to the target character set to generate a first target recognition model includes: The target character set is adjusted according to the respondent information to obtain a third character set; the respondent information includes Chinese character mastery level information and respondent experience information; the number of characters in the third character set is less than the number of characters in the target character set; The initial recognition model is optimized according to the third character set to generate the first target recognition model.

4. The method according to claim 1 or 2, characterized in that The first recognition result includes a plurality of fourth characters; After obtaining the first recognition result, the method further includes: For any fourth character among the plurality of fourth characters, calculating the confidence of the fourth character; Determining whether the confidence level is less than a confidence threshold; If the confidence level is less than a confidence threshold, marking the fourth character as a misrecognized character; A feedback report is generated according to the misrecognized characters and context information of the misrecognized characters, and the feedback report is sent to a marking end.

5. The method according to claim 2, characterized in that Calculating a plurality of matching degrees between the first character and a plurality of candidate characters in the candidate character set, comprising: Calculate the similarity between the semantic feature vector of the first character and the semantic feature vectors of multiple candidate characters; and use the similarity as the matching degree.

6. The method according to claim 1 or 2, characterized in that Before identifying the answer image to be identified by using the first object recognition model, the method further includes: Get the initial answer image; Performing denoising processing on the initial answer image to obtain a first image; the denoising processing is to replace each pixel value in the initial answer image with the median value in its neighborhood; Performing handwriting correction processing on the first image to obtain a second image; the handwriting correction processing refers to removing broken or blurred parts of character strokes in the first image; The character outline clearing process is performed on the second image to obtain a third image; the character outline clearing process refers to enhancing edge information and outline information of the second image.

7. The method according to claim 2, characterized in that The method further comprises: For any first character, calculating the word frequency of the first character in the reference answer information and extracting a semantic feature vector of the first character; determining a recognition weight of the first character based on the word frequency and context information of the first character; the context information is determined based on the semantic feature vector of the first character and the semantic feature vectors of adjacent first characters; adjusting the initial recognition model according to a plurality of recognition weights of a plurality of first characters in the first character set to generate a second target recognition model; Recognize the answer image to be recognized by the second object recognition model to obtain a second recognition result; A final recognition result is determined according to the first recognition result and the second recognition result.

8. An image recognition device for a paper marking system, characterized in that: The device comprises: An information acquisition module, configured to acquire a candidate character set and reference answer information for a target subject; the candidate character set includes all candidate characters; a target character set determination module, configured to delete some candidate characters from the candidate character set according to the reference answer information to obtain a target character set; A model adjustment module, configured to adjust the initial recognition model according to the target character set to generate a first target recognition model; The recognition module is used to recognize the answer image to be recognized through the first target recognition model to obtain a first recognition result.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the image recognition method for a paper marking system according to any one of claims 1 to 7 by executing the computer instructions.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the image recognition method for a paper marking system according to any one of claims 1 to 7.

Images