A method, apparatus, device, and storage medium for improving the accuracy of facial recognition.

Abstract

The present invention discloses a method, device, equipment and storage medium for improving the accuracy of face recognition, and belongs to the field of image processing technology. The method comprises: obtaining a correction video; according to the correction video and a pre-recorded seating chart, obtaining the effective area where the head portraits of the persons to be named in the seating chart appear in the correction video; obtaining the video to be identified, identifying the video to be identified, and obtaining the actual area of ​​the head portraits of the persons to be identified in the video to be identified; comparing the actual area with the effective area to determine whether the recognition result of the video to be identified is correct. By obtaining the effective area where the head portraits of the persons to be identified in the video to be identified appear, and then comparing the actual area where the head portraits of the persons to be identified appear with the effective area to determine whether the recognition result of the video to be identified is correct, the misjudgment rate of face recognition in automatic roll call can be effectively reduced, and the accuracy of face recognition can be improved.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a method, device, equipment and storage medium for improving the accuracy of face recognition. Background Art

[0002] At present, AI technology is increasingly being used in life, teaching, and business, among which face detection and face recognition technologies are increasingly being used in automatic roll call. The usual approach to automatic roll call is to extract and store the facial feature values ​​of all personnel, and perform face detection and recognition on the images taken by the camera when calling the roll. If a certain face is found in the image, the roll call is considered successful. If a certain face is not found in the image, the person is considered not found.

[0003] However, in the process of automatic roll call, there is often a situation where one person is mistakenly identified as another person. Therefore, it is easy to make misjudgments in the process of automatic roll call. Summary of the invention

[0004] In view of this, the purpose of the present invention is to propose a method, device, equipment and storage medium for improving the accuracy of face recognition, which can effectively reduce the misjudgment rate of face recognition in automatic roll call and improve the accuracy of face recognition.

[0005] Based on the above purpose, the present invention provides a method for improving the accuracy of face recognition, the method comprising:

[0006] Get the calibration video;

[0007] According to the correction video and the pre-recorded seating chart, obtaining the effective area where the head portraits of the persons to be called in the seating chart appear in the correction video;

[0008] Acquire a video to be identified, identify the video to be identified, and acquire an actual area of ​​a head portrait of each person to be identified in the video to be identified;

[0009] The actual area is compared with the effective area to determine whether the recognition result of the video to be recognized is correct.

[0010] Preferably, the obtaining, based on the correction video and the pre-recorded seating table, of the effective area where the head portraits of the persons to be called in the seating table appear in the correction video comprises:

[0011] According to the corrected video, a first video frame image including the head portraits of the persons to be named is obtained;

[0012] According to the first video frame picture, a first list including the head portraits, the positions of the head portraits, and the width and height of the head portraits of the persons to be named is obtained;

[0013] Associating the head portrait position of each person to be called in the first list with the person information of each person to be called in the seating chart;

[0014] According to the personnel information of each person to be called out in the seating chart, the position of the head portrait of each person to be called out in the first list, and the width and height of the head portrait, the effective area where the head portrait of each person to be called out in the seating chart appears in the video to be identified is obtained.

[0015] More preferably, associating the head portrait position of each person to be called in the first list with the personnel information of each person to be called in the seating chart includes:

[0016] Sort the ordinates of the positions of the avatars of the persons to be called out in the first list, with the first N avatars with the largest ordinates being the first row, and so on, until the row information of the avatars of all the persons to be called out is obtained;

[0017] Sort the horizontal coordinates of the positions of the portraits of the persons to be called in each row, with the smallest horizontal coordinate being in the first column, the second smallest being in the second column, and so on, until the column information of the portraits of all the persons to be called is obtained;

[0018] According to the row and column information of the avatars of the persons to be called and the row and column information in the seating table, the positions of the avatars of the persons to be called in the first list are associated with the personnel information of the persons to be called in the seating table.

[0019] More preferably, the method further comprises pre-entering a seating chart including personnel information of all persons to be called, wherein the personnel information comprises the names, student numbers and row and column information of all persons to be called.

[0020] Preferably, the method further comprises re-identifying if the recognition result of the video to be identified is incorrect.

[0021] Based on the same inventive concept, the present invention also provides a device for improving the accuracy of face recognition, the device comprising:

[0022] An acquisition module, used for acquiring a correction video;

[0023] An effective area acquisition module, for acquiring, based on the correction video and the pre-recorded seating chart, an effective area where the head portraits of the persons to be called in the seating chart appear in the correction video;

[0024] The actual area acquisition module is used to acquire the video to be identified, identify the video to be identified, and acquire the actual area of ​​the head portrait of each person to be identified in the video to be identified;

[0025] The judgment module is used to compare the actual area with the effective area to judge whether the recognition result of the video to be recognized is correct.

[0026] Preferably, the effective area acquisition module includes:

[0027] A first video frame image acquisition unit, configured to acquire a first video frame image including the head portraits of the persons to be named according to the corrected video;

[0028] A first list acquisition unit is used to acquire, according to the first video frame picture, a first list including the head portraits, head portrait positions, and head portrait widths and heights of the persons to be named;

[0029] an associating unit, configured to associate the head portrait position of each person to be called in the first list with the personnel information of each person to be called in the seating chart;

[0030] The effective area acquisition unit is used to obtain the effective area where the head portrait of each person to be called out in the seating chart appears in the video to be identified based on the personal information of each person to be called out in the seating chart, the head portrait position and the head portrait width and height of each person to be called out in the first list.

[0031] More preferably, the association unit is used for:

[0032] Sort the ordinates of the positions of the avatars of the persons to be called out in the first list, with the first N avatars with the largest ordinates being the first row, and so on, until the row information of the avatars of all the persons to be called out is obtained;

[0033] Sort the horizontal coordinates of the positions of the portraits of the persons to be called in each row, with the smallest horizontal coordinate being in the first column, the second smallest being in the second column, and so on, until the column information of the portraits of all the persons to be called is obtained;

[0034] According to the row and column information of the avatars of the persons to be called and the row and column information in the seating table, the positions of the avatars of the persons to be called in the first list are associated with the personnel information of the persons to be called in the seating table.

[0035] More preferably, the device further comprises a seating chart entry module for pre-entering a seating chart including personnel information of all persons to be called, wherein the personnel information comprises the names, student numbers and row and column information of all persons to be called.

[0036] Preferably, the device further comprises a re-identification module, which is used for re-identifying if the identification result of the video to be identified is incorrect.

[0037] Based on the same inventive concept, the present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any one of the above-described methods when executing the program.

[0038] Based on the same inventive concept, the present invention also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute any of the above methods.

[0039] From the above, it can be seen that the method, device, equipment and storage medium for improving the accuracy of face recognition provided by the present invention obtain the effective area where the head portrait of each person to be identified in the video to be identified appears, and then compare the actual area where the head portrait of each person to be identified appears with the effective area to determine whether the recognition result of the video to be identified is correct. This can effectively reduce the misjudgment rate of face recognition in automatic roll call and improve the accuracy of face recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 is a schematic diagram of a method flow of an embodiment of the present invention;

[0042] Figure 2 Yes Figure 1 Explanation of step S20;

[0043] Figure 3 Yes Figure 2 Explanation of step S203;

[0044] Figure 4 Schematic diagram of the device structure of an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

[0046] It should be noted that all expressions using "first" and "second" in the embodiments of the present invention are for distinguishing two non-identical entities with the same name or non-identical parameters. It can be seen that "first" and "second" are only for the convenience of expression and should not be understood as limitations on the embodiments of the present invention. The subsequent embodiments will not explain this one by one.

[0047] The method for improving the accuracy of face recognition provided by an embodiment of the present invention includes:

[0048] Get the calibration video;

[0049] According to the correction video and the pre-recorded seating chart, obtaining the effective area where the head portraits of the persons to be called in the seating chart appear in the correction video;

[0050] Acquire a video to be identified, identify the video to be identified, and acquire an actual area of ​​a head portrait of each person to be identified in the video to be identified;

[0051] The actual area is compared with the effective area to determine whether the recognition result of the video to be recognized is correct.

[0052] Figure 1 Schematic diagram of the method flow of an embodiment of the present invention, as shown in Figure 1 As shown, in an embodiment of a method for improving face recognition accuracy provided by the present invention, the method includes:

[0053] S10. Obtaining a corrected video;

[0054] Install a camera in the area where all the people to be called are located. The camera installation position needs to ensure the clarity of the video and the wide angle of shooting, so that the camera can shoot all the people to be called. Before shooting, remind all the people to be called to look up and sit up straight. Look at the camera, and then turn on the camera to shoot. For example, install a 4K camera above the teacher's blackboard. The wide angle of the 4K camera is 90°. The 4K camera can shoot all students in the classroom scene. Before shooting, remind students to look up, sit up straight, look at the 4K camera above the blackboard, and then turn on the camera to start shooting.

[0055] S20. According to the correction video and the pre-recorded seating chart, obtaining the effective area where the head portraits of the persons to be named in the seating chart appear in the correction video;

[0056] Before shooting the video, pre-enter a seating chart containing the personal information of each person to be called. Detect each person to be called in the corrected video obtained in step S10, and obtain the effective area where the head portrait of each person to be called in the seating chart appears in the corrected video based on the head portrait position, head portrait width and height of each person to be called, and the personal information of each person to be called in the seating chart. It should be noted that the effective area refers to the maximum range that the head of each person to be called in the seating chart can reach in the corrected video. When identifying each person to be identified in the video to be identified, if the head portrait position of the person to be identified is included in the effective area, it is considered that the identification result of the person to be identified is correct; if the head portrait position of the person to be identified is not included in the effective area, it is considered that the identification result of the person to be identified is wrong.

[0057] S30. Obtain the video to be identified, identify the video to be identified, and obtain the actual area of ​​the head portrait of each person to be identified in the video to be identified;

[0058] Obtain a video to be identified that contains all persons to be identified, extract a second video frame from the video to be identified, the second video frame contains all persons to be identified, detect the second video frame, and obtain the head portraits, head portrait width and height, and the horizontal and vertical coordinates of the area where the head portraits are located of all persons to be identified in the video to be identified. Before automatic roll call, collect the head portraits, names, student numbers and other information of all persons to be named contained in the seating chart to establish a sample library. Compare the head portraits of all persons to be identified with the head portraits of all persons to be named in the sample library, calculate the similarity between the head portraits of all persons to be identified and the head portraits of all persons to be named in the sample library, and then compare the calculated similarity with the threshold. If the calculated similarity is greater than the threshold, the person to be identified is considered to be a sample in the sample library. For example, if the similarity between the head portrait of person a to be identified and the head portrait of person A in the sample library is greater than a threshold, for example, the threshold is 0.6, then the person a to be identified is considered to be person A in the sample library, and so on, the persons corresponding to all the persons to be identified are obtained from the sample library.

[0059] By identifying the video to be identified, a second list is obtained, which includes the head portrait of each person to be identified, the width and height of the head portrait, the actual area of ​​the head portrait, the corresponding person of each person to be identified in the sample library, and the similarity between the head portrait of each person to be identified and the head portrait of the corresponding person.

[0060] S40. Compare the actual area with the effective area to determine whether the recognition result of the video to be recognized is correct;

[0061] The actual area of ​​the head portrait of each person to be identified obtained in step S30 is compared with the effective area where the head portrait of each person to be named in the seating table obtained in step S20 appears in the video to be identified, and it is determined whether the recognition result in the second list obtained in step S30 is correct. If the actual area of ​​the head portrait of the person to be identified is included in the effective area, the recognition result of the person to be identified in the second list is correct, the person to be identified is the corresponding person in the sample library recorded in the second list, and the information about the person to be identified in the second list is recorded; if the actual area of ​​the head portrait of the person to be identified is not included in the effective area, the recognition result of the person to be identified in the second list is incorrect, the person to be identified is not the corresponding person in the sample library recorded in the second list, and the information about the person to be identified in the second list is a misjudgment item. For example, in the automatic roll call of students, according to the seating chart, the seat of student Xiao Ming is in the second row and second column. According to the corrected video, the position coordinates of Xiao Ming's avatar are: horizontal coordinate X1200, vertical coordinate Y1000, Xiao Ming's avatar width is 100, and Xiao Ming's avatar height is 100. Then the effective area where Xiao Ming's avatar appears in the corrected video is the horizontal coordinate X1200~1300, and the vertical coordinate Y1000~1100; when the camera video is recognized, the recognition result is: the avatar position coordinates, that is, the actual area of ​​the avatar, are the horizontal coordinate x500, the vertical coordinate y500, the avatar width is 50, the avatar height is 50, and the avatar position is Xiao Ming; at this time, the actual area of ​​the avatar is not within the effective area, so the avatar position is not Xiao Ming, and the recognition result is incorrect, which is a misjudgment item.

[0062] Figure 2 Yes Figure 1 Explanation of step S20 in step S20, specifically, obtaining the effective area where the head portraits of the persons to be called in the seating table appear in the correction video according to the correction video and the pre-recorded seating table in step S20 includes:

[0063] S201. According to the corrected video, a first video frame image including the head portraits of the persons to be named is obtained;

[0064] A first video frame image is extracted from the corrected video obtained in step S10, wherein the first video frame image contains the head portraits of the persons to be called, and the number of the head portraits of the persons to be called in the first video frame image is the same as the number of persons to be called in the pre-entered seating chart.

[0065] S202. According to the first video frame, a first list including the head portraits, head portrait positions, and head portrait widths and heights of the persons to be named is obtained;

[0066] The first video frame image obtained in step S201 is detected to obtain a first list, wherein the first list includes the head portrait, the head portrait position, and the head portrait width and height of each person to be named.

[0067] S203. Associating the head portrait position of each person to be called in the first list with the personnel information of each person to be called in the seating chart;

[0068] The row and column information of the portraits of the persons to be called in the first list obtained in step S202 is compared with the row and column information of the persons to be called in the seating chart, and the positions of the portraits of the persons to be identified in the first list are respectively associated with the corresponding personnel information in the seating chart.

[0069] S204. According to the personnel information of each person to be called in the seating table, the position of the head portrait of each person to be called in the first list, and the width and height of the head portrait, obtain the effective area where the head portrait of each person to be called in the seating table appears in the video to be identified;

[0070] Taking the midpoint of the area where the portraits of the persons to be called out in the first list are located as the center, the areas obtained by increasing the width and height of the portraits of the persons to be called out by 2 times respectively are used as the effective areas where the portraits of the persons to be called out in the seating chart appear in the video to be identified.

[0071] Figure 3 Yes Figure 2 Specifically, associating the head portrait position of each person to be called in the first list with the personnel information of each person to be called in the seating chart in step S203 includes:

[0072] S2031. Sort the ordinates of the positions of the avatars of the persons to be called in the first list, with the first N avatars with the largest ordinates being the first row, and so on, until the row information of the avatars of all persons to be called is obtained;

[0073] Sort the ordinates of the positions of the portraits of the persons to be called out in the first list obtained in step S202, that is, sort the Y values ​​of the positions of the portraits of the persons to be called out. The Y values ​​of the positions of the portraits of the persons to be called out can be sorted in ascending order or in descending order. The first N portraits with the largest to smallest ordinates are in the first row, and so on, until the row information of the portraits of all the persons to be called out is obtained. It should be noted that N is the number of persons to be automatically called out in each row of the seating chart.

[0074] S2032. Sort the horizontal coordinates of the positions of the portraits of the persons to be called in each row, with the smallest horizontal coordinate being the first column, the second smallest being the second column, and so on, until the column information of the portraits of all persons to be called is obtained;

[0075] The horizontal coordinates of the positions of the portraits of the N persons to be called out in each row divided in step S2031 are sorted respectively, that is, the X values ​​of the positions of the portraits of the N persons to be called out in each row are sorted, and the X values ​​of the positions of the portraits of the N persons to be called out in each row can be sorted in ascending order or in descending order. The smallest horizontal coordinate is the first column, the second smallest is the second column, and so on, until the column information of the portraits of all the persons to be called out is obtained, that is, the portraits of the N persons to be called out in each row are divided into columns.

[0076] S2033. According to the row and column information of the head portraits of the personnel to be called and the row and column information in the seating table, the head portrait positions of the personnel to be called in the first list are associated with the personnel information of the personnel to be called in the seating table;

[0077] The row information of the avatars of all persons to be called is obtained through step S2031, and the column information of the avatars of all persons to be called is obtained through step S2032. The row and column information of the avatars of all persons to be called are compared with the row and column information of the persons to be called in the pre-entered seating chart, and the position of the avatar of each person to be called in the first list is associated with the personnel information of each person to be called in the pre-entered seating chart.

[0078] Optionally, it also includes pre-entering a seating chart including personnel information of all persons to be called, wherein the personnel information includes the names, student numbers and row and column information of all persons to be called;

[0079] Before shooting the video, a seating chart containing the personnel information of all persons to be called out is entered and stored, wherein the personnel information includes the names, student numbers, and row and column information of all persons to be called out.

[0080] In the embodiment of the present invention, the method further comprises re-identifying if the recognition result of the video to be identified is incorrect.

[0081] If the recognition result of the video to be recognized is incorrect, re-recognition is required. For example, if the recognition result of the person to be recognized c obtained in step S30 is incorrect, that is, the recognition result of the person to be recognized c in the second list being sample C in the sample library is incorrect, then the person to be recognized c needs to be re-recognized.

[0082] When re-identifying the person to be identified c, firstly, temporarily remove the sample C from the sample library so that the sample C will not affect the identification result again when re-identifying; then, re-identify the person to be identified c to obtain the actual area of ​​the head portrait of the person to be identified c, and compare the actual area of ​​the head portrait of the person to be identified c obtained by re-identification with the effective area obtained in step S20 to determine whether the re-identification result of the person to be identified c is correct. If the result of the re-identification of the person to be identified c is correct, then record the result of the re-identification of the person to be identified c; if the result of the re-identification of the person to be identified c is incorrect, then repeat the above steps until it is determined that the actual area of ​​the head portrait of the person to be identified c does not belong to the effective area of ​​the head portraits of all the persons to be named in the seating table determined in step S20, and the head portrait of the person to be identified c does not belong to the persons to be named in the sample library, then record the person to be identified c as a non-sample person.

[0083] Figure 4 Schematic diagram of the device structure of an embodiment of the present invention. Figure 4 As shown, an embodiment of the present invention discloses a device for improving face recognition accuracy, the device comprising:

[0084] An acquisition module, used for acquiring a correction video;

[0085] An effective area acquisition module, for acquiring, based on the correction video and the pre-recorded seating chart, an effective area where the head portraits of the persons to be called in the seating chart appear in the correction video;

[0086] The actual area acquisition module is used to acquire the video to be identified, identify the video to be identified, and acquire the actual area of ​​the head portrait of each person to be identified in the video to be identified;

[0087] The judgment module is used to compare the actual area with the effective area to judge whether the recognition result of the video to be recognized is correct.

[0088] In the embodiment of the present invention, the effective area acquisition module includes:

[0089] A first video frame image acquisition unit, configured to acquire a first video frame image including the head portraits of the persons to be named according to the corrected video;

[0090] A first list acquisition unit is used to acquire, according to the first video frame picture, a first list including the head portraits, head portrait positions, and head portrait widths and heights of the persons to be named;

[0091] an associating unit, configured to associate the head portrait position of each person to be called in the first list with the personnel information of each person to be called in the seating chart;

[0092] The effective area acquisition unit is used to obtain the effective area where the head portrait of each person to be called out in the seating chart appears in the video to be identified based on the personal information of each person to be called out in the seating chart, the head portrait position and the head portrait width and height of each person to be called out in the first list.

[0093] In the embodiment of the present invention, the association unit is used to:

[0094] Sort the ordinates of the positions of the avatars of the persons to be called out in the first list, with the first N avatars with the largest ordinates being the first row, and so on, until the row information of the avatars of all the persons to be called out is obtained;

[0095] Sort the horizontal coordinates of the positions of the head portraits of the persons to be named in each row, with the smallest horizontal coordinate being in the first column, the second smallest being in the second column, and so on, until the column information of the head portraits of all the persons to be named is obtained;

[0096] According to the row and column information of the avatars of the persons to be called and the row and column information in the seating table, the positions of the avatars of the persons to be called in the first list are associated with the personnel information of the persons to be called in the seating table.

[0097] Specifically, the device also includes a seating chart input module for pre-inputting a seating chart including personnel information of all persons to be called, wherein the personnel information includes the names, student numbers and row and column information of all persons to be called.

[0098] Optionally, the device further comprises a re-identification module, configured to re-identify if the identification result of the video to be identified is incorrect.

[0099] The apparatus of the above-mentioned embodiment is used to implement the corresponding method in the above-mentioned embodiment, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.

[0100] An embodiment of the present invention also provides a scheduled task execution device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the program, any of the scheduled task execution methods described above is implemented.

[0101] An embodiment of the present invention further provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, wherein the computer instructions are used to enable the computer to execute to implement any of the above-mentioned scheduled task execution methods.

[0102] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present disclosure (including the claims) is limited to these examples. Within the scope of the present invention, the above embodiments or technical features in different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity.

[0103] In addition, to simplify the description and discussion, and in order not to obscure the present invention, known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided figures. In addition, the devices may be shown in the form of block diagrams to avoid obscuring the present invention, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform on which the present invention will be implemented (i.e., these details should be fully within the scope of understanding of those skilled in the art). Where specific details (e.g., circuits) are set forth to describe exemplary embodiments of the present invention, it will be apparent to those skilled in the art that the present invention may be implemented without these specific details or with variations in these specific details. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0104] Although the invention has been described in conjunction with specific embodiments of the invention, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0105] The embodiments of the present invention are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for improving face recognition accuracy, characterized in that: The method comprises: Get the calibration video; According to the correction video and the pre-recorded seating chart, obtaining the effective area where the head portraits of the persons to be called in the seating chart appear in the correction video; Acquire a video to be identified, identify the video to be identified, and acquire an actual area of ​​a head portrait of each person to be identified in the video to be identified; Comparing the actual area with the effective area to determine whether the recognition result of the video to be recognized is correct; Wherein, obtaining the effective area where the head portraits of the persons to be called in the seating table appear in the correction video according to the correction video and the pre-recorded seating table includes: According to the corrected video, a first video frame image including the head portraits of the persons to be named is obtained; According to the first video frame picture, a first list including the head portraits, the positions of the head portraits, and the width and height of the head portraits of the persons to be named is obtained; Associating the head portrait position of each person to be called in the first list with the person information of each person to be called in the seating chart; According to the personnel information of each person to be called out in the seating table, the position of the head portrait of each person to be called out in the first list, and the width and height of the head portrait, obtaining the effective area where the head portrait of each person to be called out in the seating table appears in the video to be identified; The comparing the actual area with the effective area to determine whether the recognition result of the video to be recognized is correct includes: In response to determining that the actual area of ​​the head portrait of the person to be identified is included in the valid area, the identification result of the person to be identified is correct; or in response to determining that the actual area of ​​the head portrait of the person to be identified is not included in the valid area, the identification result of the person to be identified is incorrect.

2. The method for improving face recognition accuracy as claimed in claim 1, characterized in that: The step of associating the head portrait position of each person to be called in the first list with the person information of each person to be called in the seating chart includes: Sort the ordinates of the positions of the avatars of the persons to be called out in the first list, with the first N avatars with the largest ordinates being the first row, and so on, until the row information of the avatars of all the persons to be called out is obtained; Sort the horizontal coordinates of the positions of the portraits of the persons to be called in each row, with the smallest horizontal coordinate being in the first column, the second smallest being in the second column, and so on, until the column information of the portraits of all the persons to be called is obtained; According to the row and column information of the avatars of the persons to be called and the row and column information in the seating table, the positions of the avatars of the persons to be called in the first list are associated with the personnel information of the persons to be called in the seating table.

3. The method for improving face recognition accuracy as claimed in claim 1, characterized in that: The method also includes pre-entering a seating chart including personnel information of all persons to be called, wherein the personnel information includes the names, student numbers and row and column information of all persons to be called.

4. The method for improving face recognition accuracy as claimed in claim 1, characterized in that: The method further includes re-identifying if the identification result of the video to be identified is incorrect.

5. A device for improving the accuracy of face recognition, characterized in that: The device comprises: An acquisition module, used for acquiring a correction video; An effective area acquisition module, for acquiring, based on the correction video and the pre-recorded seating chart, an effective area where the head portraits of the persons to be called in the seating chart appear in the correction video; An actual area acquisition module is used to acquire a video to be identified, identify the video to be identified, and acquire an actual area of ​​the head portrait of each person to be identified in the video to be identified; A judgment module, used for comparing the actual area with the effective area to judge whether the recognition result of the video to be recognized is correct; the comparing the actual area with the effective area to judge whether the recognition result of the video to be recognized is correct comprises: in response to determining that the actual area of ​​the head portrait of the person to be recognized is included in the effective area, judging that the recognition result of the person to be recognized is correct; or in response to determining that the actual area of ​​the head portrait of the person to be recognized is not included in the effective area, judging that the recognition result of the person to be recognized is incorrect; Wherein, the effective area acquisition module includes: A first video frame image acquisition unit, configured to acquire a first video frame image including the head portraits of the persons to be named according to the corrected video; A first list acquisition unit is used to acquire, according to the first video frame picture, a first list including the head portraits, head portrait positions, and head portrait widths and heights of the persons to be named; an associating unit, configured to associate the head portrait position of each person to be called in the first list with the personnel information of each person to be called in the seating chart; The effective area acquisition unit is used to obtain the effective area where the head portrait of each person to be called out in the seating chart appears in the video to be identified based on the personal information of each person to be called out in the seating chart, the head portrait position and the head portrait width and height of each person to be called out in the first list.

6. The device for improving face recognition accuracy as claimed in claim 5, characterized in that: The association unit is used to: Sort the ordinates of the positions of the avatars of the persons to be called out in the first list, with the first N avatars with the largest ordinates being the first row, and so on, until the row information of the avatars of all the persons to be called out is obtained; Sort the horizontal coordinates of the positions of the portraits of the persons to be called in each row, with the smallest horizontal coordinate being in the first column, the second smallest being in the second column, and so on, until the column information of the portraits of all the persons to be called is obtained; According to the row and column information of the avatars of the persons to be called and the row and column information in the seating table, the positions of the avatars of the persons to be called in the first list are associated with the personnel information of the persons to be called in the seating table.

7. The device for improving face recognition accuracy as claimed in claim 5, characterized in that: The device also includes a seating chart input module, which is used to pre-input a seating chart including personnel information of all persons to be called, wherein the personnel information includes the names, student numbers and row and column information of all persons to be called.

8. The device for improving face recognition accuracy as claimed in claim 5, characterized in that: The device also includes a re-identification module, which is used for re-identifying if the identification result of the video to be identified is incorrect.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method according to any one of claims 1 to 4 is implemented.

10. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to cause the computer to execute any one of claims 1 to 4.

Images