Answer sheet fill-in recognition method, system and related apparatus

Abstract

The application discloses a kind of answer sheet filling recognition method, system and related device, the method includes: image acquisition is carried out to answer sheet, obtains the option image corresponding to multiple candidate options;Wherein each described option image includes filling area;The contour information corresponding to the option image is obtained;Wherein, the contour information is determined based on the option image and the multiple image sub-block corresponding to the option image;Based on the contour information, the filling state information of the candidate option corresponding to the filling area is obtained.Through the above mode, the accuracy of the answer sheet filling recognition can be improved.

Description

Technical Field

[0001] This application relates to the field of image recognition technology, and in particular to a method, system and related apparatus for recognizing answer sheet filling. Background Technology

[0002] Answer sheet marking recognition technology, as an important component of modern examination scoring systems, has a development history closely linked to the advancement of information technology. Current automatic recognition methods typically involve scanning completed answer sheets with a scanner and determining whether they are marked or not based on a threshold parameter. However, the accuracy of threshold-based methods is relatively low because the resulting image is affected by many factors, such as the ink density on the answer sheet paper, the scanner's imaging environment, and subjective factors like the student's marking habits.

[0003] Therefore, improving the accuracy of answer sheet marking and recognition has become an urgent problem to be solved. Summary of the Invention

[0004] The main technical problem addressed by this application is to provide a method, system, and related apparatus for answer sheet filling and recognition, which can improve the accuracy of answer sheet filling and recognition.

[0005] To address the aforementioned technical problems, this application provides a method for recognizing answer sheet filling, comprising: acquiring an image of an answer sheet to obtain option images corresponding to multiple candidate options; wherein each option image includes a filling area; obtaining contour information corresponding to the option image; wherein the contour information is determined based on the option image and multiple image sub-blocks corresponding to the option image; and obtaining filling status information of the candidate option corresponding to the filling area based on the contour information.

[0006] To address the aforementioned technical problems, another technical solution adopted in this application is: providing an answer sheet filling and recognition system, comprising: a data acquisition module for acquiring images of the answer sheet to obtain option images corresponding to multiple candidate options; wherein each option image includes a filling area; an acquisition module for acquiring contour information corresponding to the option image; wherein the contour information is determined based on the option image and multiple image sub-blocks corresponding to the option image; and a processing module for acquiring the filling status information of the candidate option corresponding to the filling area based on the contour information.

[0007] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide an electronic device, including a memory and a processor coupled to each other, wherein the memory stores program instructions, and the processor is used to execute the program instructions to implement the method mentioned in the above technical solution.

[0008] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide a computer-readable storage medium having program instructions stored thereon, wherein the program instructions, when executed by a processor, implement the method mentioned in the above technical solution.

[0009] The beneficial effects of this application are as follows: Unlike existing technologies, the answer sheet filling and recognition method proposed in this application obtains the option images corresponding to each candidate option in the answer sheet, and determines the corresponding contours based on the option images and multiple image sub-blocks obtained by segmenting the option images. By combining the contours corresponding to the option images and image sub-blocks, as well as the width of the filling area, the filling status information of the candidate options is determined, thereby improving the accuracy of answer sheet filling and recognition. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0011] Figure 1 This is a flowchart illustrating one implementation method of the answer sheet filling and recognition method of this application;

[0012] Figure 2 yes Figure 1 The flowchart of step S102 corresponds to another embodiment;

[0013] Figure 3 yes Figure 2 A schematic diagram of step S202 corresponding to one embodiment;

[0014] Figure 4 yes Figure 2 The flowchart of step S203 corresponds to another embodiment;

[0015] Figure 5 yes Figure 1 The flowchart of step S103 corresponds to another embodiment;

[0016] Figure 6 yes Figure 1 A flowchart of an embodiment following step S103;

[0017] Figure 7 This is a flowchart illustrating one implementation method for obtaining the ratio of the filled area;

[0018] Figure 8 yes Figure 1A flowchart of another implementation method following step S103;

[0019] Figure 9 This is a schematic diagram of one embodiment of the answer sheet filling and recognition system of this application;

[0020] Figure 10 This is a schematic diagram of the structure of one embodiment of the electronic device of this application;

[0021] Figure 11 This is a schematic diagram of one embodiment of the computer-readable storage medium of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments, and different embodiments can be adaptively combined. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0023] Please see Figure 1 , Figure 1 This is a flowchart illustrating one embodiment of the answer sheet filling and recognition method of this application. The method includes:

[0024] S101: Capture images of the answer sheet to obtain option images corresponding to multiple candidate options; where each option image includes a fill-in area.

[0025] In one embodiment, to recognize the content filled in on the answer sheet, a scanning device is used to acquire an image of the answer sheet to obtain a corresponding initial image. Based on the initial image, the option image corresponding to each candidate option on the answer sheet is determined, and each option image includes a corresponding filling area.

[0026] Specifically, the initial image is divided in advance based on the arrangement information of each candidate option on the answer sheet to obtain the option image corresponding to each candidate option.

[0027] S102: Obtain the contour information corresponding to the option image; wherein, the contour information is determined based on the option image and multiple image sub-blocks corresponding to the option image.

[0028] In one embodiment, each option image is segmented to obtain multiple corresponding image sub-blocks. Based on the complete option image and each corresponding image sub-block, contour information is extracted, which is used at least to characterize the boundaries of the filled content in the option image and the image sub-block.

[0029] Specifically, contours are extracted from the option image and each corresponding image sub-block, and the extracted contours of the option image and each image sub-block are used as contour information.

[0030] S103: Based on the contour information, obtain the filling status information of the candidate options corresponding to the filling area.

[0031] In one embodiment, the filling status information of the corresponding candidate option is judged based on the extracted contour information to determine whether the corresponding candidate option is in a filled state or an unfilled state.

[0032] Specifically, based on the contour information of the option image and each image sub-block, the maximum contour width corresponding to each option image and each image sub-block is determined. The maximum contour width is compared with the width of the filled area, and based on the comparison result, the filling status of the candidate option corresponding to each filled area is determined to be either filled or unfilled.

[0033] The answer sheet filling and recognition method proposed in this application acquires the option images corresponding to each candidate option in the answer sheet, and determines the corresponding contours based on the option images and multiple image sub-blocks obtained by segmenting the option images. By combining the contours corresponding to the option images and image sub-blocks, as well as the width of the filling area, the filling status information of the candidate options is determined, thereby improving the accuracy of answer sheet filling and recognition.

[0034] Please see Figure 2 , Figure 2 yes Figure 1 The flowchart of step S102 corresponds to another embodiment. Specifically, the implementation process of step S102 includes:

[0035] S201: Obtain the first edge contour corresponding to the option image, and obtain the corresponding reference length based on the first edge contour.

[0036] In one embodiment, contour recognition is performed on the complete option image to obtain the corresponding first edge contour. The total length of the first edge contour is used as a reference length.

[0037] It should be noted that when the corresponding area to be filled in the option image is not filled, the first edge contour is the boundary of the corresponding area to be filled, and the total length of the boundary of the area to be filled is used as the reference length.

[0038] In another embodiment, to avoid the reference length value of the first edge contour being too large due to irregular filling, which would interfere with subsequent judgment, a pre-set length threshold is obtained. When the reference length corresponding to the option image is greater than the length threshold, the length threshold is used as the corresponding reference length.

[0039] Specifically, the aforementioned length threshold is determined based on the total length of the boundary of the filled area. For example, the total length of the boundary of the filled area is used as the length threshold. When the reference length of the first edge contour corresponding to the complete option image is greater than the length threshold, the length threshold is used as the reference length of the corresponding option image.

[0040] Alternatively, a pre-set expansion threshold can be obtained, and the sum of the total length of the filled area boundary and the expansion threshold can be used as the length threshold. When the reference length of the first edge contour corresponding to the complete option image is greater than the length threshold, the length threshold is used as the reference length of the corresponding option image. The specific data of the above expansion threshold can be set according to actual needs.

[0041] S202: Divide the option image into multiple image sub-blocks.

[0042] In one embodiment, the option image is divided to obtain multiple corresponding image sub-blocks.

[0043] In one implementation scenario, please refer to Figure 3 , Figure 3 yes Figure 2 The schematic diagram of step S202 corresponds to one embodiment. For example... Figure 3 As shown, based on the option image corresponding to candidate option "A", the corresponding horizontal and vertical midlines are determined. The option image is then divided according to the horizontal midline X and the vertical midline Y to obtain four corresponding image sub-blocks.

[0044] S203: Based on the maximum contour width corresponding to at least some image sub-blocks, obtain the corresponding reference width and value.

[0045] In one embodiment, the maximum contour widths corresponding to at least some of the image sub-blocks are summed to obtain the corresponding reference width and value.

[0046] In one implementation scenario, please refer to [the relevant documentation]. Figure 3 The option image is divided into multiple image sub-blocks along the horizontal and vertical midlines. The sum of the maximum contour widths corresponding to at least one image sub-block is used as the reference width sum value.

[0047] Please see Figure 4 , Figure 4 yes Figure 2 The flowchart of step S203 corresponds to another embodiment. Specifically, the implementation process of step S203 includes:

[0048] S301: For the image sub-block located on the first side of the vertical center line, the maximum value of the corresponding maximum contour width is used as the first reference width.

[0049] In one embodiment, the option image is divided into multiple image sub-blocks along the horizontal and vertical midlines. For the image sub-block located on the first side of the vertical midline, the maximum value among the corresponding maximum contour widths is used as the first reference width.

[0050] For a specific application scenario, please refer to [link / reference]. Figure 3 Two image sub-blocks, namely image sub-block 1 and image sub-block 2, are distributed to the left of the vertical midline Y. The maximum value of the maximum contour width corresponding to image sub-block 1 and image sub-block 2 is used as the first reference width.

[0051] S302: For the image sub-block located on the second side of the vertical center line, the maximum value of the corresponding maximum contour width is used as the second reference width.

[0052] In one embodiment, for an image sub-block located on the second side of the vertical midline, the maximum value of the corresponding maximum contour width is used as the second reference width.

[0053] For a specific application scenario, please refer to [link / reference]. Figure 3 Two image sub-blocks, namely image sub-block 3 and image sub-block 4, are distributed to the right of the vertical center line Y. The maximum value of the maximum contour width corresponding to image sub-block 3 and image sub-block 4 is used as the second reference width.

[0054] S303: Use the sum of the first reference width and the second reference width as the reference width sum value.

[0055] In one embodiment, the obtained first reference width and second reference width are added together to obtain a reference width sum value. This reference width sum value is used to characterize the lateral width of the content to be filled in the corresponding filling area in the horizontal direction.

[0056] The above scheme determines a first reference width based on the image sub-block on the first side of the vertical center line and a second reference width based on the image sub-block on the second side of the vertical center line. The obtained reference width and value represent the maximum width of the filled content projected in the horizontal direction, which helps to determine the filling status information of the corresponding candidate option based on the reference width and value and the width of the filled area. This improves the efficiency and accuracy of answer sheet filling recognition.

[0057] In another embodiment, the method of dividing the option image into multiple image sub-blocks can also be different. Specifically, a horizontal center line and multiple vertical dividing lines are determined for the option image, which are used to divide the option image equally in the vertical direction. For example, a horizontal center line and two vertical dividing lines are determined for the option image, and the option image is evenly divided into six image sub-blocks by the horizontal center line and the two dividing lines. That is, the two dividing lines evenly divide the option image into three image columns, and the horizontal center line evenly divides each image column into two image sub-blocks, one above the other. Further, for the two image sub-blocks corresponding to each image column, the maximum value of the corresponding maximum contour width is used as the reference width. The sum of the reference widths corresponding to all image columns is used as the reference width sum.

[0058] Please see Figure 5 , Figure 5 yes Figure 1 The flowchart of step S103 corresponds to another embodiment. Specifically, the implementation process of step S103 includes:

[0059] S401: Use the maximum value among all the maximum contour widths corresponding to the option image as the third reference width.

[0060] In one embodiment, in response to obtaining the maximum contour width corresponding to each image sub-block in the option image through the corresponding implementation described above, the maximum value among all the maximum contour widths corresponding to the option image is used as the third reference width. Determining the maximum third reference width provides a basis for judging subsequent filling status information and helps improve the accuracy of the judgment.

[0061] S402: In response to a first ratio of the third reference width to the width of the filled area being greater than a first threshold, or a second ratio of the sum of the reference widths to the width of the filled area being greater than a second threshold, and a third ratio of the reference length to the area of ​​the filled area being greater than a third threshold, the filling status information of the candidate option is determined to be a filling status.

[0062] In one embodiment, a first ratio of a third reference width to the width of the filled area, a second ratio of the sum of reference widths to the width of the filled area, and a third ratio of the reference length to the area of ​​the filled area are obtained. Provided that the third ratio is greater than a third threshold, if the first ratio is greater than the first threshold or the second ratio is greater than the second threshold, the filling status information of the candidate option is determined to be in a filled state; otherwise, the filling status information of the candidate option is determined to be in an unfilled state. The aforementioned first threshold, second threshold, and third threshold can be estimated by those skilled in the art, or they can be derived through multiple experiments.

[0063] In another embodiment, to further improve recognition efficiency, the filling status information of the candidate option can be determined solely based on the first ratio and the third ratio. For example, when the first ratio is greater than a first threshold and the third ratio is greater than a third threshold, the filling status information of the candidate option is determined to be filled; otherwise, the filling status information of the candidate option is determined to be unfilled.

[0064] Alternatively, the filling status information of the candidate option can be determined solely based on the second ratio and the third ratio. That is, when the second ratio is greater than the first threshold and the third ratio is greater than the third threshold, the filling status information of the candidate option is determined to be filled; otherwise, the filling status information of the candidate option is determined to be unfilled.

[0065] Please see Figure 6 , Figure 6 yes Figure 1 The flowchart following step S103 corresponds to one implementation method. Specifically, in response to the fact that the filling status information of all candidate options matching the current title is in an unfilled state, the process after step S103 further includes:

[0066] S501: For each option image corresponding to the current title, obtain the corresponding reference data.

[0067] In one embodiment, multiple titles are identified from an initial image obtained from the answer sheet, and option images for multiple candidate options corresponding to each title are determined. For each option image corresponding to the current title, its corresponding reference data is obtained.

[0068] Specifically, the reference data for each option image includes a reference grayscale, a filling area ratio, and a fourth reference width. The reference grayscale is determined based on the grayscale values ​​of each initial pixel in the central region of the corresponding option image. The filling area ratio is determined based on the filling content of the corresponding filling area. The fourth reference width is determined based on the contour information in the corresponding option image.

[0069] In a specific application scenario, each title corresponds to candidate option "A", candidate option "B" and candidate option "C". Then, the reference gray level, filling area ratio and fourth reference width corresponding to each of the above candidate options are determined respectively.

[0070] S502: Sort the reference data corresponding to the current title according to the preset sorting rules to obtain the reference sequence.

[0071] In one embodiment, for each candidate option corresponding to the current title, the corresponding reference data is sorted according to a preset sorting rule to obtain a corresponding reference sequence. The preset sorting rule can be sorting by numerical values ​​from largest to smallest.

[0072] In one implementation scenario, in response to the reference data corresponding to each option image including reference grayscale, filling area ratio, and fourth reference width, the reference grayscale of each option image corresponding to the current title is sorted according to a preset sorting rule to obtain the corresponding first reference sequence; the filling area ratio of each option image corresponding to the current title is sorted according to a preset sorting rule to obtain the corresponding second reference sequence; and the fourth reference width of each option image corresponding to the current title is sorted according to a preset sorting rule to obtain the corresponding third reference sequence.

[0073] In another implementation, the reference data corresponding to each candidate option can also be sorted in ascending order of numerical values ​​to obtain the corresponding reference sequence.

[0074] S503: Obtain the first data and the second data in the reference sequence, and update the filling status information of the candidate options based on the ratio of the first data and the second data and the candidate options corresponding to the first data.

[0075] In one embodiment, the reference data corresponding to each option image includes a reference grayscale, a filling area ratio, and a fourth reference width. The reference data with the largest value in the reference sequence is used as the first data, and the reference data with the second largest value is used as the second data. The filling status information of the corresponding candidate option is updated based on the ratio of the first data and the second data, and the candidate option corresponding to the first data.

[0076] In one implementation scenario, when the reference data for each option image includes reference grayscale, fill area ratio, and fourth reference width, the reference ratios corresponding to the reference grayscale, fill area ratio, and fourth reference width are obtained respectively. The reference ratio is the ratio of the corresponding first data to the corresponding second data.

[0077] Specifically, based on the first reference sequence, corresponding first data and second data are determined, and their ratio is used as the first reference ratio; based on the second reference sequence, corresponding first data and second data are determined, and their ratio is used as the second reference ratio; and based on the third reference sequence, corresponding first data and second data are determined, and their ratio is used as the third reference ratio.

[0078] Furthermore, in response to all reference ratios satisfying the first preset update condition and all candidate options corresponding to the first data being the same, the filling status information of the candidate options corresponding to the first data is updated to the filling status.

[0079] Specifically, when the first reference ratio obtained above is greater than the fourth threshold, the second reference ratio is greater than the fifth threshold, and the third reference ratio is less than the sixth threshold, it is determined that the first preset update condition is met. And when all candidate options corresponding to the first data obtained from the first, second, and third reference sequences are the same, it is determined that the second preset update condition is met. Furthermore, when both the first and second preset update conditions are met simultaneously, the filling status information of any candidate option corresponding to the first data under the current title is updated to a filled status.

[0080] Optionally, in other embodiments, to save computational resources, the reference data corresponding to each option image may also include at least one of reference grayscale, filling area ratio, and fourth reference width. For example, the reference data corresponding to each option image includes reference grayscale and fourth reference width. The reference grayscale values ​​of each option image corresponding to the current title are sorted according to a preset sorting rule to obtain a corresponding reference sequence; and the fourth reference widths of each option image corresponding to the current title are sorted according to a preset sorting rule to obtain a corresponding reference sequence. For each reference sequence, the reference data with the largest value is taken as the first data, and the reference data with the second largest value is taken as the second data. Based on the ratio of the first data and the second data, and the candidate option corresponding to the first data, the filling status information of the corresponding candidate option is updated. The specific update process can be referred to the corresponding embodiment described above.

[0081] Please see Figure 7 , Figure 7 This is a flowchart illustrating one implementation method for obtaining the filling area ratio. Specifically, the process for obtaining the filling area ratio includes:

[0082] S601: Obtain multiple reference pixels corresponding to the boundary of the filling area, and determine the grayscale threshold based on the grayscale values ​​of the multiple reference pixels.

[0083] In one embodiment, the initial pixel located at the boundary of the filling area is used as a reference pixel. The grayscale value corresponding to each reference pixel is obtained, and a grayscale threshold is determined based on the grayscale value corresponding to the reference pixel.

[0084] Specifically, the median of the grayscale values ​​of all reference pixels is obtained, and this median is multiplied by a preset weight parameter to obtain the grayscale threshold. The aforementioned weight parameter can be estimated by relevant technical personnel, or it can be derived through multiple experiments.

[0085] In another embodiment, the average of the gray values ​​of all reference pixels is obtained, and the average is multiplied by the weight parameter to obtain the gray threshold.

[0086] S602: Using a grayscale threshold, the option image is binarized to obtain a binarized image corresponding to the option image; wherein, the binarized image includes multiple filled pixels.

[0087] In one embodiment, for each initial pixel in the option image, it is determined whether the grayscale value corresponding to the initial pixel is greater than or equal to a grayscale threshold. If it is greater than or equal to a grayscale threshold, the grayscale value of the initial pixel is updated to a first value. If it is less than a first value, the grayscale value of the initial pixel is updated to a second value. The first value is greater than the second value.

[0088] In a specific application scenario, the first value mentioned above is 255, and the second value mentioned above is 0.

[0089] S603: Obtain the filling area based on all filling pixels of the option image.

[0090] In one embodiment, for the binarized image corresponding to the obtained option image, pixels with grayscale values ​​of a first value are used as fill pixels, and the corresponding fill area is determined based on the number of fill pixels. This fill area is used to characterize the size of the fill content in the option image.

[0091] S604: The ratio of the filled area to the filled area is used as the filling area ratio.

[0092] In one embodiment, the ratio of the filled area to the area of ​​the filled region is obtained, and this ratio is used as the filling area ratio.

[0093] The above scheme improves the accuracy of judgment by obtaining the proportion of the filled area and combining different types of data during the judgment of the filling status information.

[0094] Please see Figure 8 , Figure 8 yes Figure 1 The flowchart following step S103 corresponds to another implementation method. Specifically, after step S103, the method further includes:

[0095] S701: Based on the filling status information and reference data corresponding to the candidate options, obtain the corresponding confidence score.

[0096] In one embodiment, a pre-set reference weight is obtained, and the corresponding confidence score is calculated based on the reference weight, the filling status information corresponding to the candidate option, and the reference data corresponding to the candidate option.

[0097] In one implementation scenario, the reference data includes reference grayscale, fill area ratio, and fourth reference width. Pre-set first reference weights, second reference weights, and third reference weights are obtained. A first difference between a first value and the reference grayscale is calculated, and the ratio between the first difference and the first value is calculated. The first sub-score is obtained by multiplying this ratio by the first reference weight. The second sub-score is obtained by multiplying the fill area ratio by the second reference weight. The third sub-score is obtained by multiplying the fourth reference width by the option image width and the third reference weight. The specific calculation formulas for the first, second, and third sub-scores are as follows:

[0098]

[0099] R a =R×W g

[0100]

[0101] Among them, R g G represents the score of the first sub-sub. avg W represents the reference grayscale. g R represents the first reference weight; a R represents the second sub-score, R represents the percentage of the area filled, and W represents the percentage of the area filled. g R represents the second reference weight; l W represents the score for the third child. maxf W represents the fourth reference width, and w represents the width of the option image. l This represents the third reference weight. The first, second, and third reference weights mentioned above were obtained by relevant technical personnel through estimation, or they may have been derived through multiple experiments.

[0102] Furthermore, the confidence score is determined based on the filling status information, first sub-score, second sub-score, and third sub-score corresponding to the candidate options.

[0103] Specifically, the product of the first value, the first sub-score, the second sub-score, and the third sub-score is used as the reference product. If the candidate option's corresponding fill status information is "filled," the aforementioned reference product is used as the confidence score matching the corresponding candidate option. If the candidate option's corresponding fill status information is "unfilled," the difference between the first value and the reference product is used as the confidence score matching the corresponding candidate option. The specific formula for calculating the confidence score is as follows:

[0104] R f =255×R l ×R a ×R g

[0105] R b =255-255×R l ×R a ×R g

[0106] Among them, R f This represents the confidence score for the corresponding candidate option when the form is filled in; R b This indicates the confidence score for the corresponding candidate option when the field is unfilled.

[0107] In another embodiment, the first reference weight, the second reference weight, and the third reference weight may also be determined based on the border information of the filled area. The border information includes the color depth and border length of the border. The darker the border color, the greater the first reference weight, the second reference weight, and the third reference weight. The smaller the border length, the greater the first reference weight, the second reference weight, and the third reference weight.

[0108] S702: Send the option images corresponding to confidence scores that are less than the scoring threshold to the human reviewer.

[0109] In one implementation, it is determined whether the confidence score corresponding to the candidate option is less than a preset score threshold. If it is less, the option image corresponding to the candidate option is sent to a human for review. If it is greater than or equal to the score threshold, the judgment of the filling status information of the candidate option is considered accurate, and no review is required.

[0110] In another implementation, if the confidence score of a candidate option is less than the score threshold, then all option images corresponding to the title of that candidate option are sent to a human for review.

[0111] The above scheme further improves the accuracy of recognition by determining a confidence score to further verify the filling status information of each candidate option.

[0112] Please see Figure 9 , Figure 9 This is a schematic diagram of one embodiment of the answer sheet filling and recognition system of this application. The answer sheet filling and recognition system includes a data acquisition module 10, an acquisition module 20, and a processing module 30 that are coupled to each other.

[0113] The acquisition module 10 is used to acquire images of the answer sheet and obtain option images corresponding to multiple candidate options; wherein, each option image includes a filled area.

[0114] The acquisition module 20 is used to acquire the contour information corresponding to the option image; wherein, the contour information is determined based on the option image and multiple image sub-blocks corresponding to the option image.

[0115] The processing module 30 is used to obtain the filling status information of the candidate options corresponding to the filling area based on the contour information.

[0116] In one embodiment, the acquisition module 20 acquires the contour information corresponding to the option image, including: acquiring a first edge contour corresponding to the option image, acquiring a corresponding reference length based on the first edge contour; dividing the option image into multiple image sub-blocks; acquiring a second edge contour corresponding to each image sub-block, determining the maximum contour width corresponding to each second edge contour; and acquiring a corresponding reference width and value based on the maximum contour width corresponding to at least some of the image sub-blocks.

[0117] In one embodiment, the option image corresponds to a horizontal centerline and a vertical centerline. The acquisition module 20 acquires a corresponding reference width and value based on the maximum contour width corresponding to at least some image sub-blocks, including: for image sub-blocks located on the first side of the vertical centerline, taking the maximum value of the corresponding maximum contour width as a first reference width; and for image sub-blocks located on the second side of the vertical centerline, taking the maximum value of the corresponding maximum contour width as a second reference width; and taking the sum of the first reference width and the second reference width as the reference width sum value.

[0118] In one embodiment, the processing module 30 obtains the filling status information of the corresponding candidate option based on the contour information, including: taking the maximum value among all the maximum contour widths corresponding to the option image as the third reference width; in response to a first ratio of the third reference width to the width of the filling area being greater than a first threshold or a second ratio of the sum of the reference widths to the width of the filling area being greater than a second threshold, and a third ratio of the reference length to the area of ​​the filling area being greater than a third threshold, determining the filling status information of the candidate option as a filling status.

[0119] In one implementation, please refer to... Figure 9 The answer sheet filling recognition system proposed in this application also includes an update module 40 coupled to the processing module 30. Responding to the fact that the filling status information of all candidate options matched with the current title is unfilled, after obtaining the filling status information of the candidate options corresponding to the filled area based on the contour information, the update module 40 is used to obtain corresponding reference data for each option image corresponding to the current title; sort the reference data corresponding to the current title according to a preset sorting rule to obtain a reference sequence; obtain the first data and the second data in the reference sequence; and update the filling status information of the candidate options based on the ratio of the first data and the second data, and the candidate options corresponding to the first data.

[0120] In one embodiment, the reference data of the option image includes a corresponding reference grayscale, a filling area ratio, and a fourth reference width. First data and second data in the reference sequence are obtained. The update module 40 updates the filling status information of the candidate options based on the ratio of the first data and the second data, and the candidate options corresponding to the first data. This includes: obtaining reference ratios corresponding to the reference grayscale, the filling area ratio, and the fourth reference width; wherein the reference ratio is the ratio of the corresponding first data to the corresponding second data; and in response to all reference ratios satisfying a first preset update condition and all candidate options corresponding to the first data being the same, updating the filling status information of the candidate options corresponding to the first data to a filling status.

[0121] In one embodiment, the reference grayscale is determined based on the grayscale values ​​corresponding to each initial pixel in the central region of the option image, and the fourth reference width is determined based on the contour information in the option image. The step of the update module 40 obtaining the filling area ratio includes: obtaining multiple reference pixels corresponding to the boundary of the filling area; determining a grayscale threshold based on the grayscale values ​​of the multiple reference pixels; performing binarization processing on the option image using the grayscale threshold to obtain a binarized image corresponding to the option image; wherein the binarized image includes multiple filling pixels; obtaining the filling area based on all the filling pixels of the option image; and using the ratio of the filling area to the area of ​​the filling region as the filling area ratio.

[0122] In one implementation, please refer to... Figure 9 The answer sheet filling recognition system proposed in this application also includes a verification module 50 coupled to the update module 40. After obtaining the filling status information of the candidate options corresponding to the filling area based on the contour information, the verification module 50 is used to obtain the corresponding confidence score based on the filling status information of the candidate options and the reference data; and send the option image corresponding to the confidence score that is less than the score threshold to the human end for verification.

[0123] Please see Figure 10 , Figure 10This is a schematic diagram of one embodiment of the electronic device of this application. The electronic device includes a memory 60 and a processor 70 coupled to each other. The memory 60 stores program instructions, and the processor 70 executes the program instructions to implement the methods mentioned in any of the above embodiments. Specifically, the electronic device includes, but is not limited to, desktop computers, laptops, tablets, servers, etc., and is not limited thereto. In addition, the processor 70 may also be called a CPU (Center Processing Unit). The processor 70 may be an integrated circuit chip with signal processing capabilities. The processor 70 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor may be a microprocessor or any conventional processor. In addition, the processor 70 may be implemented by integrated circuit chips.

[0124] Please see Figure 11 , Figure 11 This is a schematic diagram of a computer-readable storage medium according to an embodiment of the present application. The computer-readable storage medium 80 stores program instructions 90 that can be executed by a processor. When the program instructions 90 are executed by the processor, they implement the methods mentioned in any of the above embodiments.

[0125] In the several embodiments provided in this application, it should be understood that the disclosed methods and apparatus can be implemented in other ways. For example, the apparatus implementations described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0126] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0127] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0128] If the integrated unit is implemented as 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 this application, in essence, 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. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0129] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for recognizing answer sheet filling, characterized in that, include: The answer sheet is image captured to obtain an option image corresponding to each candidate option; wherein each option image includes a fill-in area; Obtain the contour information corresponding to the option image; wherein the contour information is determined based on the option image and multiple image sub-blocks corresponding to the option image; Based on the contour information, obtain the filling status information of the candidate options corresponding to the filling area; In response to the fact that the filling status information of all candidate options matching the current title is in an unfilled state, for each option image corresponding to the current title, corresponding reference data is obtained; the reference data corresponding to the current title is sorted according to a preset sorting rule to obtain a reference sequence; the first data and the second data in the reference sequence are obtained, and the filling status information of the candidate options is updated based on the ratio of the first data and the second data and the candidate option corresponding to the first data; wherein, the first data is the reference data with the largest value in the reference sequence, and the second data is the reference data with the second largest value in the corresponding reference sequence; The process of updating the filling status information of the candidate option includes: obtaining reference ratios corresponding to the reference grayscale, the filling area ratio, and the fourth reference width, wherein the fourth reference width is determined based on the contour information in the corresponding option image; wherein the reference ratio is the ratio of the first data to the second data; in response to all the reference ratios satisfying the first preset update condition and all the candidate options corresponding to the first data being the same, updating the filling status information of the candidate option corresponding to the first data to the filling status; The process of obtaining the filling status information of the candidate option corresponding to the filling area based on the contour information further includes: obtaining a corresponding confidence score based on the filling status information and reference data corresponding to the candidate option; sending the option image corresponding to the confidence score that is less than the score threshold to a human reviewer; wherein each option image corresponds to the reference data, and the reference data includes the corresponding reference grayscale, filling area ratio and fourth reference width.

2. The method according to claim 1, characterized in that, The step of obtaining the contour information corresponding to the option image includes: Obtain the first edge contour corresponding to the option image, and based on the total length of the first edge contour, obtain the corresponding reference length; and, The option image is divided into multiple image sub-blocks; Obtain the second edge contour corresponding to each image sub-block, and determine the maximum contour width corresponding to each second edge contour; Based on the maximum contour width corresponding to at least a portion of the image sub-blocks, obtain the corresponding reference width and value.

3. The method according to claim 2, characterized in that, The option image corresponds to a horizontal center line and a vertical center line. Obtaining the corresponding reference width and value based on the maximum contour width corresponding to at least a portion of the image sub-blocks includes: For the image sub-block located on the first side of the vertical midline, the maximum value among the corresponding maximum contour widths is used as the first reference width; and, For the image sub-block located on the second side of the vertical midline, the maximum value among the corresponding maximum contour widths is used as the second reference width; The sum of the first reference width and the second reference width is used as the reference width sum value.

4. The method according to claim 2, characterized in that, The step of obtaining the filling status information of the candidate options corresponding to the filling area based on the contour information includes: The maximum value among all the maximum contour widths corresponding to the option image is used as the third reference width; In response to a first ratio of the third reference width to the width of the filled area being greater than a first threshold, or a second ratio of the sum of the reference widths to the width of the filled area being greater than a second threshold, and a third ratio of the reference length to the area of ​​the filled area being greater than a third threshold, the filling status information of the candidate option is determined to be a filling status.

5. The method according to claim 1, characterized in that, The reference grayscale is determined based on the grayscale values ​​corresponding to each initial pixel in the central region of the option image. The step of obtaining the filling area ratio includes: Obtain multiple reference pixels corresponding to the boundary of the filling area, and determine a grayscale threshold based on the grayscale values ​​of the multiple reference pixels; Using the grayscale threshold, the option image is binarized to obtain a binarized image corresponding to the option image; wherein, the binarized image includes multiple filled pixels; Based on all the filled pixels of the option image, obtain the filled area; The ratio of the filled area to the area of ​​the filled region is used as the filling area ratio.

6. A system for recognizing and filling out answer sheets, characterized in that, include: The acquisition module is used to acquire images of the answer sheet to obtain an option image corresponding to each candidate option; wherein each option image includes a filled area; An acquisition module is used to acquire contour information corresponding to the option image; wherein the contour information is determined based on the option image and multiple image sub-blocks corresponding to the option image; The processing module is used to obtain the filling status information of the candidate option corresponding to the filling area based on the contour information; In response to the fact that the filling status information of all candidate options matching the current title is in an unfilled state, for each option image corresponding to the current title, corresponding reference data is obtained; the reference data corresponding to the current title is sorted according to a preset sorting rule to obtain a reference sequence; the first data and the second data in the reference sequence are obtained, and the filling status information of the candidate options is updated based on the ratio of the first data and the second data and the candidate option corresponding to the first data; wherein, the first data is the reference data with the largest value in the reference sequence, and the second data is the reference data with the second largest value in the corresponding reference sequence; The process of updating the filling status information of the candidate option includes: obtaining reference ratios corresponding to the reference grayscale, the filling area ratio, and the fourth reference width, wherein the fourth reference width is determined based on the contour information in the corresponding option image; wherein the reference ratio is the ratio of the first data to the second data; in response to all the reference ratios satisfying the first preset update condition and all the candidate options corresponding to the first data being the same, updating the filling status information of the candidate option corresponding to the first data to the filling status; The process of obtaining the filling status information of the candidate option corresponding to the filling area based on the contour information further includes: obtaining a corresponding confidence score based on the filling status information and reference data corresponding to the candidate option; sending the option image corresponding to the confidence score that is less than the score threshold to a human reviewer; wherein each option image corresponds to the reference data, and the reference data includes the corresponding reference grayscale, filling area ratio and fourth reference width.

7. An electronic device, characterized in that, include: A memory and a processor are coupled to each other, the memory storing program instructions, and the processor executing the program instructions to implement the method as described in any one of claims 1-5.

8. A computer-readable storage medium having program instructions stored thereon, characterized in that, When the program instructions are executed by the processor, they implement the method as described in any one of claims 1-5.

Images