Identity recognition method and device, electronic equipment and storage medium

Abstract

The invention relates to an identity recognition method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the recognition processing of image observation information of at least one to-be-recognized region, and obtaining a track recognition result in each to-be-recognized region, the track identification result comprises time information and first attribute information of at least one employee track; according to second attribute information of at least one to-be-identified employee, third attribute information of the at least one to-be-identified area and a track identification result in each to-be-identified area, a probability value of each to-be-identified employee appearing in each to-be-identified area is determined, and the second attribute information comprises an employee identity; and matching the at least one to-be-identified area with the at least one to-be-identified employee according to the track identification result in each to-be-identified area and the probability value of each to-be-identified employee appearing in each to-be-identified area, and obtaining the number and identity of the employees in each to-be-identified area.

Description

Technical Field

[0001] The present disclosure relates to the field of computer vision technology, and in particular to an identity recognition method, device, electronic device and storage medium. Background Art

[0002] In recent years, with the rapid development of artificial intelligence technology, all walks of life have benefited from the convenience brought by artificial intelligence technology. Computer vision technology can automatically identify target information, such as identifying employees within an enterprise, thereby freeing users from manual identification and reducing the labor intensity of users. However, in related technologies, when using computer vision technology to identify employees, the recognition accuracy is low. Summary of the invention

[0003] The present disclosure provides an identity recognition method, device, equipment and storage medium to solve the defects in the related art.

[0004] According to a first aspect of an embodiment of the present disclosure, there is provided an identity recognition method, the method comprising:

[0005] Performing recognition processing on the image observation information of at least one area to be recognized, and obtaining a trajectory recognition result in each area to be recognized, wherein the trajectory recognition result includes time information and first attribute information of at least one employee's trajectory;

[0006] Determine a probability value of each employee to be identified appearing in each area to be identified according to the second attribute information of at least one employee to be identified, the third attribute information of the at least one area to be identified, and the trajectory identification result in each area to be identified, wherein the second attribute information includes the identity of the employee;

[0007] According to the trajectory recognition result in each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, the at least one area to be identified and the at least one employee to be identified are matched to obtain the number of employees and employee identities in each area to be identified.

[0008] In one embodiment of the present disclosure, determining the probability value of each employee to be identified appearing in each area to be identified based on the second attribute information of at least one employee to be identified, the third attribute information of at least one area to be identified, and the trajectory recognition result in each area to be identified, includes:

[0009] Determining a matching degree between each employee to be identified and each area to be identified based on the second attribute information of the at least one employee to be identified and the third attribute information of the at least one area to be identified;

[0010] According to the matching degree between each employee to be identified and each area to be identified, and the trajectory identification result in each area to be identified, the probability value of each employee to be identified appearing in each area to be identified is determined.

[0011] In one embodiment of the present disclosure, determining the matching degree between each employee to be identified and each area to be identified based on the second attribute information of the at least one employee to be identified and the third attribute information of the at least one area to be identified includes:

[0012] The matching degree between each employee to be identified and each area to be identified is determined according to the second attribute information of the at least one employee to be identified, the third attribute information of the at least one area to be identified, and the shift information.

[0013] In one embodiment of the present disclosure, the second attribute information includes employee skills, and the third attribute information includes skills required by the region.

[0014] In one embodiment of the present disclosure, the first attribute information includes gender information;

[0015] The method of determining the probability value of each employee to be identified appearing in each area to be identified according to the matching degree between each employee to be identified and each area to be identified, and the trajectory identification result in each area to be identified, comprises:

[0016] According to the trajectory recognition results in each area to be identified, the working hours of employees of each gender in each area to be identified are determined accordingly;

[0017] According to the matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified, the probability value of each employee to be identified appearing in each area to be identified is determined.

[0018] In one embodiment of the present disclosure, determining the probability value of each employee to be identified appearing in each area to be identified according to the matching degree between each employee to be identified and each area to be identified and the working hours of employees of each gender in each area to be identified includes:

[0019] The matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified, are input into the prediction model to obtain the probability value of each employee to be identified appearing in each area to be identified output by the prediction model.

[0020] In one embodiment of the present disclosure, the prediction model is obtained through training, and the training includes:

[0021] Performing recognition processing on historical image observation information of at least one work area to obtain trajectory recognition results in each work area, wherein the historical image observation information of each work area is marked with employee identities;

[0022] Determining a matching degree between each employee identity and each work area according to the second attribute information corresponding to each employee identity annotated by the historical image observation information of the at least one work area and the third attribute information of the at least one work area;

[0023] Determine the working hours of employees of each gender in each work area according to the trajectory recognition result in each work area;

[0024] The prediction model is trained using the matching degree between each employee identity and each work area, the working hours of employees of each gender in each work area of ​​the at least one work area, and the employee identity annotated with the historical image observation information of each work area of ​​the at least one work area.

[0025] In one embodiment of the present disclosure, the at least one to-be-identified area is matched with the at least one to-be-identified employee according to the trajectory recognition result in each to-be-identified area and the probability value of each to-be-identified employee appearing in each to-be-identified area to obtain the number of employees and employee identities in each to-be-identified area, including:

[0026] For each area to be identified, clustering and merging the at least one employee trajectory according to the time information and the first attribute information of at least one employee trajectory in the trajectory recognition result in the area to be identified, to obtain an employee recognition result in the area to be identified, wherein the employee recognition result includes the time information and the first attribute information of at least one employee;

[0027] According to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, the at least one area to be identified and the at least one employee to be identified are matched to obtain the number and identity of employees in each area to be identified.

[0028] In one embodiment of the present disclosure, the method further includes:

[0029] Determine the working frequency of each employee according to the duration of the image observation information and the time information of each employee in the employee recognition result, wherein the working frequency includes the ratio of the time information of the employee to the image observation information;

[0030] Delete employees whose working frequency is lower than a preset threshold in the employee identification results.

[0031] In one embodiment of the present disclosure, the first attribute information includes the type of work clothes;

[0032] The method further comprises:

[0033] Delete employees whose uniform type is not the type to be identified in the employee identification results.

[0034] In one embodiment of the present disclosure, matching the at least one area to be identified and the at least one employee to be identified according to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, and obtaining the number of employees and employee identities in each area to be identified, includes:

[0035] According to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, a greedy matching algorithm is used to match the at least one area to be identified and the at least one employee to be identified to obtain the number of employees and employee identities in each area to be identified.

[0036] In one embodiment of the present disclosure, matching the at least one area to be identified and the at least one employee to be identified according to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, and obtaining the number of employees and employee identities in each area to be identified, includes:

[0037] Matching the at least one area to be identified and the at least one employee to be identified according to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified;

[0038] In response to a successful one-to-one matching between the at least one area to be identified and the at least one employee to be identified, the number of employees and employee identities in each area to be identified according to the matching result;

[0039] In response to the at least one area to be identified and the at least one employee to be identified not being successfully matched one-to-one, the at least one area to be identified and the at least one employee to be identified are matched according to the time information of at least one employee in the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, so as to obtain the number of employees and employee identities in each area to be identified.

[0040] In one embodiment of the present disclosure, the method further includes:

[0041] The at least one employee to be identified is determined according to the attendance record.

[0042] According to a second aspect of an embodiment of the present disclosure, there is provided an identity recognition device, the device comprising:

[0043] An identification module, used to perform identification processing on the image observation information of at least one area to be identified, and obtain a trajectory identification result in each area to be identified, wherein the trajectory identification result includes time information and first attribute information of at least one employee's trajectory;

[0044] a prediction module, configured to determine a probability value of each to-be-identified employee appearing in each to-be-identified area according to second attribute information of at least one to-be-identified employee, third attribute information of at least one to-be-identified area, and a trajectory recognition result in each to-be-identified area, wherein the second attribute information includes an employee identity;

[0045] The matching module is used to match the at least one area to be identified and the at least one employee to be identified according to the trajectory recognition result in each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, so as to obtain the number of employees and employee identities in each area to be identified.

[0046] In one embodiment of the present disclosure, the prediction module is specifically used to:

[0047] Determining a matching degree between each employee to be identified and each area to be identified based on the second attribute information of the at least one employee to be identified and the third attribute information of the at least one area to be identified;

[0048] According to the matching degree between each employee to be identified and each area to be identified, and the trajectory identification result in each area to be identified, the probability value of each employee to be identified appearing in each area to be identified is determined.

[0049] In one embodiment of the present disclosure, the prediction module is used to determine the matching degree between each employee to be identified and each area to be identified based on the second attribute information of the at least one employee to be identified and the third attribute information of the at least one area to be identified, specifically for:

[0050] The matching degree between each employee to be identified and each area to be identified is determined according to the second attribute information of the at least one employee to be identified, the third attribute information of the at least one area to be identified, and the shift information.

[0051] In one embodiment of the present disclosure, the second attribute information includes employee skills, and the third attribute information includes skills required by the region.

[0052] In one embodiment of the present disclosure, the first attribute information includes gender information;

[0053] The method of determining the probability value of each employee to be identified appearing in each area to be identified according to the matching degree between each employee to be identified and each area to be identified, and the trajectory identification result in each area to be identified, comprises:

[0054] According to the trajectory recognition results in each area to be identified, the working hours of employees of each gender in each area to be identified are determined accordingly;

[0055] According to the matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified, the probability value of each employee to be identified appearing in each area to be identified is determined.

[0056] In one embodiment of the present disclosure, the prediction module is used to determine the probability value of each employee to be identified appearing in each area to be identified based on the matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified, and is specifically used to:

[0057] The matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified, are input into the prediction model to obtain the probability value of each employee to be identified appearing in each area to be identified output by the prediction model.

[0058] In one embodiment of the present disclosure, the device further includes a training module, which is used to:

[0059] Performing recognition processing on historical image observation information of at least one work area to obtain trajectory recognition results in each work area, wherein the historical image observation information of each work area is marked with employee identities;

[0060] Determining a matching degree between each employee identity and each work area according to the second attribute information corresponding to each employee identity annotated by the historical image observation information of the at least one work area and the third attribute information of the at least one work area;

[0061] Determine the working hours of employees of each gender in each work area according to the trajectory recognition result in each work area;

[0062] The prediction model is trained using the matching degree between each employee identity and each work area, the working hours of employees of each gender in each work area of ​​the at least one work area, and the employee identity annotated with the historical image observation information of each work area of ​​the at least one work area.

[0063] In one embodiment of the present disclosure, the matching module is specifically used for:

[0064] For each area to be identified, clustering and merging the at least one employee trajectory according to the time information and the first attribute information of at least one employee trajectory in the trajectory recognition result in the area to be identified, to obtain an employee recognition result in the area to be identified, wherein the employee recognition result includes the time information and the first attribute information of at least one employee;

[0065] According to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, the at least one area to be identified and the at least one employee to be identified are matched to obtain the number and identity of employees in each area to be identified.

[0066] In one embodiment of the present disclosure, the device further includes a deletion module, configured to:

[0067] Determine the working frequency of each employee according to the duration of the image observation information and the time information of each employee in the employee recognition result, wherein the working frequency includes the ratio of the time information of the employee to the image observation information;

[0068] Delete employees whose working frequency is lower than a preset threshold in the employee identification results.

[0069] In one embodiment of the present disclosure, the first attribute information includes the type of work clothes;

[0070] The device also includes a deletion module, which is specifically used to:

[0071] Delete employees whose uniform type is not the type to be identified in the employee identification results.

[0072] In one embodiment of the present disclosure, the matching module is used to match the at least one area to be identified and the at least one employee to be identified according to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, and when obtaining the number of employees and employee identities in each area to be identified, specifically to:

[0073] According to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, a greedy matching algorithm is used to match the at least one area to be identified and the at least one employee to be identified to obtain the number of employees and employee identities in each area to be identified.

[0074] In one embodiment of the present disclosure, the matching module is used to match the at least one area to be identified and the at least one employee to be identified according to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, and when obtaining the number of employees and employee identities in each area to be identified, specifically to:

[0075] Matching the at least one area to be identified and the at least one employee to be identified according to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified;

[0076] In response to a successful one-to-one matching between the at least one area to be identified and the at least one employee to be identified, the number of employees and employee identities in each area to be identified according to the matching result;

[0077] In response to the at least one area to be identified and the at least one employee to be identified not being successfully matched one-to-one, the at least one area to be identified and the at least one employee to be identified are matched according to the time information of at least one employee in the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, so as to obtain the number of employees and employee identities in each area to be identified.

[0078] In one embodiment of the present disclosure, the device further includes an attendance module, which is used to:

[0079] The at least one employee to be identified is determined according to the attendance record.

[0080] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, the device comprising a memory and a processor, the memory being used to store computer instructions executable on the processor, the processor being used to implement the method described in the first aspect when executing the computer instructions.

[0081] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the method described in the first aspect is implemented.

[0082] According to the above embodiments, the identity recognition method provided by the present disclosure can perform recognition processing on the image observation information of at least one area to be recognized, and obtain the trajectory recognition result in each area to be recognized; determine the probability value of each employee to be recognized appearing in each area to be recognized according to the second attribute information of at least one employee to be recognized, the third attribute information of the at least one area to be recognized, and the trajectory recognition result; match the at least one area to be recognized and the at least one employee to be recognized according to the trajectory recognition result in each area to be recognized and the probability value of each employee to be recognized appearing in each area to be recognized, and obtain the number of employees and employee identities in each area to be recognized. The method uses computational visual recognition technology to recognize the first attribute information with low accuracy, and further combines the non-visual information of the employee to be recognized and the area to be recognized to obtain the identity of the employee in each area to be recognized, thereby improving the recognition accuracy of the employee identity without improving the recognition accuracy of computer vision.

[0083] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0085] Figure 1 is a flow chart of an identity recognition method shown in an embodiment of the present disclosure;

[0086] Figure 2 is a flow chart of a probability value prediction method shown in one embodiment of the present disclosure;

[0087] Figure 3 is a schematic diagram of the construction and application of a prediction model shown in an embodiment of the present disclosure;

[0088] Figure 4 is a flow chart of a matching method shown in an embodiment of the present disclosure;

[0089] Figure 5 is a flow chart of an identity recognition method shown in an embodiment of the present disclosure;

[0090] Figure 6 is a structural schematic diagram of an identity recognition device shown in an embodiment of the present disclosure;

[0091] Figure 7 It is a schematic diagram of the structure of an electronic device shown in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0092] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0093] The terms used in this disclosure are for the purpose of describing specific embodiments only and are not intended to limit the disclosure. The singular forms of "a", "said" and "the" used in this disclosure and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0094] It should be understood that although the terms first, second, third, etc. may be used in the present disclosure to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present disclosure, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0095] In recent years, with the rapid development of artificial intelligence technology, all walks of life have benefited from the convenience brought by artificial intelligence technology. Computer vision technology can automatically identify target information, such as identifying employees within an enterprise, thereby freeing users from manual identification and reducing the labor intensity of users. However, in related technologies, when using computer vision technology to identify employees, the identity of employees is often determined by performing face recognition on images or videos containing employees; when the distance between the employee and the camera that collects the image or video is far, the depth of the employee in the image or video will be large, and employees wearing hats, masks, etc. will cause certain occlusion of the employee's head, especially the face. These situations will lead to reduced accuracy of face recognition, that is, inaccurate identification of the employee's identity.

[0096] Based on this, at least one embodiment of the present disclosure provides an identity recognition method, which can be used to identify employees in each work area in a work environment, and the recognition results can be used for manpower scheduling, attendance statistics, performance appraisal, etc. Exemplarily, the method can perform computer vision recognition based on image observation information, and combine non-visual information of employees and the work environment to ultimately determine the identity and number of employees in each area.

[0097] Please refer to the attached Figure 1, which exemplarily shows the process of the method, including steps S101 to S103.

[0098] In step S101, image observation information of at least one area to be identified is identified and processed to obtain a trajectory identification result in each area to be identified, wherein the trajectory identification result includes time information and first attribute information of at least one employee trajectory.

[0099] The working environment is divided into multiple working areas, and the working areas to be identified by the method are the areas to be identified. For example, if there are employees working in a certain working area, the working area can be used as the area to be identified, and if there are no employees working in a certain working area, the working area can be not used as the area to be identified.

[0100] At least one observation device (such as a camera, etc.) is provided in the working environment, and each observation device can collect observation videos for at least one working area. The image observation information can be the observation video collected by the observation device (such as a camera) for the area to be identified, and the observation video covers the area to be identified. The image observation information of different areas to be identified can be independent of each other or integrated together; for example, if the observation video collected by a certain observation device covers multiple areas to be identified, the image observation information of the multiple areas to be identified are integrated together.

[0101] The employee trajectory may be the trajectory formed by the employee's activities in the observation video. It is understandable that since an employee may frequently enter and exit the area to be identified, different employee trajectories may belong to the same employee. The time information of the employee trajectory may include the time when the employee enters the observation video (i.e., the starting point of the trajectory) and the time when the employee leaves the observation video (i.e., the end point of the trajectory); the time in the observation video may be represented by a sequence of image frames in the video, for example, in a certain scene, the employee enters the observation video at the 50th frame of the observation video and leaves the observation video at the 120th frame. The first attribute information is certain attribute information of the employee to whom the employee trajectory belongs. This attribute information may be information with a lower accuracy than the employee's identity, such as the employee's gender and work uniform type. The work uniform type may include a type to be identified (e.g., employee uniform) and a type not to be identified (e.g., non-employee uniform, etc.); since the accuracy of the first attribute information is lower than that of the employee's identity (i.e., the employee's facial information), the recognition accuracy of the first attribute information is higher and the requirements for the recognition technology are lower.

[0102] Exemplarily, computer vision (CV) technology can be used to identify image observation information to obtain trajectory recognition results in each area to be identified. For example, a CV module can be configured, and the CV module includes a human figure detection reasoning unit and a work clothes recognition reasoning unit; the human figure detection reasoning unit can perform human figure detection on each frame of the observed video to obtain the position, gender, orientation and occlusion information of the human figure, and track each human figure between different frames. When a human figure leaves the observed video, an employee trajectory is formed; for each employee trajectory, the number of image frames of male gender and the number of image frames of female gender are counted respectively, and the gender with more image frames is determined as the gender of the employee trajectory; from multiple image frames with correct gender recognition, an image frame with the closest orientation to the positive direction and the least occlusion is selected, and the work clothes recognition reasoning unit is ordered to identify this image frame to obtain the work clothes type.

[0103] In step S102, the probability value of each employee to be identified appearing in each area to be identified is determined based on the second attribute information of at least one employee to be identified, the third attribute information of at least one area to be identified, and the trajectory recognition result in each area to be identified, wherein the second attribute information includes the identity of the employee.

[0104] The employees to be identified may be part or all of the employees. Before this step, at least one employee to be identified may be determined based on the attendance records. For example, all employees who have clocked in and not clocked out when the method is executed may be determined as employees to be identified.

[0105] The second attribute information of the employee to be identified may be non-visual information, i.e., information that does not need to be identified in step S101; for example, the second attribute information may include employee identity (e.g., number or name, etc.), employee gender, employee skills, etc. The third attribute information of the area to be identified may be non-visual information, i.e., information that does not need to be identified in step S101; for example, the third attribute information may include skills required for the area, etc.

[0106] In one possible embodiment, the following Figure 2 This step is performed in the manner shown, including sub-steps S1021 and S1022.

[0107] In sub-step S1021, the matching degree between each employee to be identified and each area to be identified is determined according to the second attribute information of at least one employee to be identified and the third attribute information of at least one area to be identified.

[0108] When the second attribute information includes employee skills and the third attribute information includes regional required skills, this sub-step can determine the matching degree of each employee to be identified and each area to be identified based on the employee skills of each employee to be identified in at least one employee to be identified and the regional required skills of each area to be identified in at least one area to be identified.

[0109] The matching degree between each employee to be identified and each area to be identified can be determined in combination with the current scheduling information, that is, the matching degree between each employee to be identified and each area to be identified can be determined based on the second attribute information of at least one employee to be identified, the third attribute information of at least one area to be identified, and the scheduling information, wherein the scheduling information may include the employees scheduled in each area to be identified. For example, the matching degree between each employee to be identified and each area to be identified can be determined based on the employee skills of each employee to be identified in at least one employee to be identified, the area required skills of each area to be identified in at least one area to be identified, and the scheduling information.

[0110] For example, a matching degree determination rule may be configured, such as if the employee A to be identified is arranged in the area a to be identified in the shift information, then the matching degree between the employee A to be identified and the area a to be identified is 2; if the employee A to be identified is not arranged in the area a to be identified in the shift information, and the employee skills of the employee A to be identified are the same as the area required skills of the area a to be identified, then the matching degree between the employee A to be identified and the area a to be identified is 1; if the employee A to be identified is not arranged in the area a to be identified in the shift information, and the employee skills of the employee A to be identified are different from the area required skills of the area a to be identified, then the matching degree between the employee A to be identified and the area a to be identified is 0. The matching degree between each employee to be identified and each area to be identified may be determined based on the employee skills of each employee to be identified in at least one employee to be identified, the area required skills of each area to be identified in at least one area to be identified, the shift information, and the above rules.

[0111] In sub-step S1022, the probability value of each employee to be identified appearing in each area to be identified is determined according to the matching degree between each employee to be identified and each area to be identified, and the trajectory identification result in each area to be identified.

[0112] Exemplarily, according to the trajectory recognition results in each area to be identified, the working hours of employees of each gender in each area to be identified are correspondingly determined. For example, for the trajectory recognition results in each area to be identified, the duration of all male employee trajectories can be summed to obtain the working hours of male employees, and the duration of all female employee trajectories can be summed to obtain the working hours of female employees. This sub-step can determine the probability value of each employee to be identified appearing in each area to be identified based on the matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified.

[0113] For example, the prediction model can be trained in the following manner to enable it to have the function of predicting the appearance of a certain employee in a certain work area, wherein the prediction model can select a lightGBM model, a GBDT (Gradient Boosting Decision Tree) model, an xgboost model, etc.

[0114] Please refer to the attached Figure 3 , which takes the xgboost model as an example, exemplarily illustrates the process of building a model and using the model to predict the probability of each employee to be identified appearing in each area to be identified.

[0115] The historical image observation information of at least one work area is identified and processed to obtain the trajectory identification result in each work area, wherein the historical image observation information of each work area is marked with the identity of the employee. The historical image observation information is the image observation information of the work area in a certain period of time, and the number of employees and the identity of the employees (i.e., the identity of each employee) in the work area during the period of time are known and have been marked on the historical image observation information; it can be understood that the historical image observation information is the same as the details of the image observation information in step S101, which will not be repeated here, and the identification processing of the historical image observation information is the same as the details of the identification processing of the image observation information in step S101, which will not be repeated here.

[0116] The matching degree between each employee identity and each work area is determined based on the second attribute information corresponding to each employee identity annotated by the historical image observation information of at least one work area and the third attribute information of at least one work area. It can be understood that the method of determining the matching degree in this step can be the same as the method of determining the matching degree in sub-step S1021, which will not be repeated here.

[0117] According to the trajectory recognition results in each work area in at least one work area, the working hours of employees of each gender in each work area are correspondingly determined. It can be understood that the method of determining the working hours of employees of each gender in this step can be the same as the method of determining the working hours of employees of each gender in sub-step S1022, and will not be repeated here.

[0118] The prediction model is trained using the matching degree between each employee identity and each work area, the working hours of employees of each gender in each work area in at least one work area, and the employee identity annotated with the historical image observation information of each work area in at least one work area. For example, the matching degree between each employee identity and each work area, the working hours of employees of each gender in each work area in at least one work area are used as independent variables, and the employee identity annotated with the historical image observation information of each work area in at least one work area is used as a dependent variable to construct and train a prediction model such as an xgboost model.

[0119] Based on the prediction model constructed and trained as described above, sub-step S1023 can input the matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified, into the prediction model to obtain the probability value of each employee to be identified appearing in each area to be identified output by the prediction model. In step S103, at least one area to be identified and at least one employee to be identified are matched according to the trajectory recognition results in each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, to obtain the number of employees and employee identities in each area to be identified.

[0120] In one possible embodiment, the following Figure 4 This step is performed in the manner shown, including sub-steps S1031 and S1032.

[0121] In sub-step 1031, for each area to be identified, at least one employee trajectory is clustered and merged according to the time information and the first attribute information of at least one employee trajectory in the trajectory recognition result in the area to be identified, so as to obtain the employee recognition result in the area to be identified, wherein the employee recognition result includes the time information and the first attribute information of at least one employee.

[0122] At least one employee determined in this step is an employee to be identified in the work area, that is, an employee whose identity is to be confirmed. Exemplarily, in the trajectory recognition result, employee trajectories with non-overlapping time and the same first attribute information (for example, the same gender and the same work uniform type) can be merged into the recognition result of a certain employee, and the employee's time information (that is, the union of the time information of each employee trajectory corresponding to the employee) and the first attribute information (that is, the first attribute information of the employee trajectory corresponding to the employee) can be determined.

[0123] It is understandable that there may be some employees passing by in the area to be identified. These employees passing by are not the employees whose identities are to be confirmed in the area to be identified. These employees passing by have one characteristic: they are in the area to be identified for a short time. Therefore, the working frequency of each employee can be determined according to the duration of the image observation information and the time information of each employee in the employee identification result, wherein the working frequency includes the ratio of the employee's time information to the image observation information; and the employees whose working frequency is lower than the preset threshold in the employee identification result are deleted. Thus, the interference factors in the area to be identified are removed, and it can also avoid that a certain employee exists in the employee identification results of different areas to be identified. For example, if an employee works in area A and passes by area B during work, the employee exists in the employee identification results of both areas A and B, and the interference of the employee can be removed from the employee identification result of area B through the above correction method.

[0124] It is understandable that not all employees in the area to be identified need to be identified. For example, non-employees and visitors do not need to be identified, because the purpose of identity identification is often human resource scheduling and performance appraisal. Therefore, employees whose uniform types are not those to be identified in the employee identification results can be deleted. For example, if the type to be identified is employee type, then the non-employee employees in the employee identification results can be deleted. This removes interference factors in the area to be identified, and can also prevent a certain employee from existing in the employee identification results of different areas to be identified. For example, if a non-employee person is active in area A and in area B, then the non-employee person exists in the employee identification results of areas A and B. The interference of the employee can be removed from the employee identification results of areas A and B through the above correction method.

[0125] In sub-step 1032, according to the employee identification results of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, at least one area to be identified and at least one employee to be identified are matched to obtain the number of employees and employee identities in each area to be identified.

[0126] Exemplarily, at least one area to be identified is matched with at least one employee to be identified based on the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified. For example, a greedy matching algorithm is used to match at least one area to be identified with at least one employee to be identified based on the principle that the area has the highest probability of appearing and an employee can only appear in one area.

[0127] In response to the successful one-to-one matching of at least one area to be identified and at least one employee to be identified, the number of employees and the identity of employees in each area to be identified are obtained according to the matching results; in response to the unsuccessful one-to-one matching of at least one area to be identified and at least one employee to be identified, at least one area to be identified and at least one employee to be identified are matched according to the time information of at least one employee in the employee identification results of each area to be identified, and the probability value of each employee to be identified appearing in each area to be identified, to obtain the number of employees and the identity of employees in each area to be identified. In other words, if the match is not successful, the constraint condition is relaxed to successfully match. For example, when the first attribute information is gender information, the gender constraint can be relaxed to successfully match when the match is not successful, so as to ensure the success rate of matching and improve the success rate of identity recognition.

[0128] See attached Figure 5 , which exemplarily shows a flow chart of an identity recognition method obtained by combining the above multiple embodiments. In this example, the identity information of employees in three working areas of a certain enterprise, namely, area A, area B, and area C, is recognized within 10 minutes. The following is a detailed description of this example.

[0129] The image observation information of the three to-be-identified areas, A, B, and C, is identified using the CV vision algorithm to obtain the trajectory recognition results in each to-be-identified area. Taking area A as an example, its trajectory recognition results are shown in Table 1 below (each row in the table represents a trajectory):

[0130] Table 1: Trajectory recognition results in area A

[0131]

[0132]

[0133] Next, the employee gender, skill information, employee attendance information, required skills for each work area, and shift information are input into the xgboost regional prediction model to obtain the probability value of each employee to be identified appearing in each area to be identified as shown in Table 2 below.

[0134] Table 2: Probability value table

[0135]

[0136]

[0137] According to the trajectory recognition results shown in Table 1 and the probability values ​​shown in Table 2, greedy matching is performed on at least one area to be identified and at least one employee to be identified to obtain the number of employees and employee identities in each area to be identified.

[0138] In this step, multiple trajectories in Table 1 are merged to obtain the employee recognition results shown in Table 3 below:

[0139] Table 3: Employee identification results

[0140] Work Area Work clothes gender Frequency of occurrence in the work area A Employee Group male 0.75 A Employee Group male 0.52 B Employee Group female 0.28 B Employee Group female 0.33 C Non-employee group male 0.48 C Non-employee group male 0.41 C Employee Group female 0.3

[0141] Delete the employees whose frequency is lower than 0.3 and whose uniforms are in the non-employee group in Table 3, and match the remaining employees with the employee area records with higher probability values ​​in Table 2, and obtain the matching results shown in Table 4 below (i.e., the number of employees and employee identities (employee codes) in each area to be identified):

[0142] Table 4: Matching results table

[0143] Work Area Employee code C 5 B 1 B 5 A 2 A 3

[0144] According to the above embodiments, the identity recognition method provided by the present disclosure can perform recognition processing on the image observation information of at least one area to be recognized, and obtain the trajectory recognition result in each area to be recognized; determine the probability value of each employee to be recognized appearing in each area to be recognized according to the second attribute information of at least one employee to be recognized, the third attribute information of at least one area to be recognized, and the trajectory recognition result; match at least one area to be recognized and at least one employee to be recognized according to the trajectory recognition result in each area to be recognized and the probability value of each employee to be recognized appearing in each area to be recognized, and obtain the number of employees and employee identities in each area to be recognized. The method uses computational visual recognition technology to recognize the first attribute information with low accuracy, and combines the non-visual information of the employee to be recognized and the area to be recognized to obtain the identity of the employee in each area to be recognized, thereby improving the recognition accuracy of the employee identity without improving the recognition accuracy of computer vision.

[0145] According to a second aspect of the embodiment of the present disclosure, an identity recognition device is provided. Figure 6 , the device comprises:

[0146] The recognition module 601 is used to perform recognition processing on the image observation information of at least one to-be-recognized area to obtain a trajectory recognition result in each to-be-recognized area, wherein the trajectory recognition result includes time information and first attribute information of at least one employee trajectory;

[0147] A prediction module 602 is used to determine a probability value of each to-be-identified employee appearing in each to-be-identified area according to second attribute information of at least one to-be-identified employee, third attribute information of at least one to-be-identified area, and a trajectory recognition result in each to-be-identified area, wherein the second attribute information includes an employee identity;

[0148] The matching module 603 is used to match at least one area to be identified and at least one employee to be identified according to the trajectory recognition results in each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, so as to obtain the number and identity of employees in each area to be identified.

[0149] In one embodiment of the present disclosure, the prediction module is specifically used to:

[0150] Determining a matching degree between each employee to be identified and each area to be identified based on the second attribute information of at least one employee to be identified and the third attribute information of at least one area to be identified;

[0151] According to the matching degree between each employee to be identified and each area to be identified, and the trajectory identification result in each area to be identified, the probability value of each employee to be identified appearing in each area to be identified is determined.

[0152] In one embodiment of the present disclosure, the prediction module is used to determine the matching degree between each employee to be identified and each area to be identified based on the second attribute information of at least one employee to be identified and the third attribute information of at least one area to be identified, specifically for:

[0153] The matching degree between each employee to be identified and each area to be identified is determined according to the second attribute information of at least one employee to be identified, the third attribute information of at least one area to be identified, and the shift information.

[0154] In one embodiment of the present disclosure, the second attribute information includes employee skills, and the third attribute information includes skills required by the region.

[0155] In one embodiment of the present disclosure, the first attribute information includes gender information;

[0156] According to the matching degree between each employee to be identified and each area to be identified, and the trajectory identification result in each area to be identified, the probability value of each employee to be identified appearing in each area to be identified is determined, including:

[0157] According to the trajectory recognition results in each area to be identified, the working hours of employees of each gender in each area to be identified are determined accordingly;

[0158] According to the matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified, the probability value of each employee to be identified appearing in each area to be identified is determined.

[0159] In one embodiment of the present disclosure, the prediction module is used to determine the probability value of each employee to be identified appearing in each area to be identified based on the matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified, and is specifically used to:

[0160] The matching degree between each employee to be identified and each area to be identified, as well as the working hours of employees of each gender in each area to be identified, are input into the prediction model to obtain the probability value of each employee to be identified appearing in each area to be identified output by the prediction model.

[0161] In one embodiment of the present disclosure, the apparatus further includes a training module, which is used to:

[0162] Performing recognition processing on historical image observation information of at least one work area to obtain trajectory recognition results in each work area, wherein the historical image observation information of each work area is marked with employee identities;

[0163] Determine the matching degree between each employee identity and each work area according to the second attribute information corresponding to each employee identity marked by the historical image observation information of at least one work area and the third attribute information of at least one work area;

[0164] According to the trajectory recognition results in each work area of ​​at least one work area, correspondingly determining the working hours of employees of each gender in each work area;

[0165] The prediction model is trained using the matching degree between each employee identity and each work area, the working hours of employees of each gender in each work area in at least one work area, and the employee identities annotated with the historical image observation information of each work area in at least one work area.

[0166] In one embodiment of the present disclosure, the matching module is specifically used for:

[0167] For each area to be identified, clustering and merging at least one employee trajectory according to the time information and the first attribute information of at least one employee trajectory in the trajectory recognition result in the area to be identified, to obtain an employee recognition result in the area to be identified, wherein the employee recognition result includes the time information and the first attribute information of at least one employee;

[0168] According to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, at least one area to be identified and at least one employee to be identified are matched to obtain the number and identity of employees in each area to be identified.

[0169] In one embodiment of the present disclosure, the device further includes a deletion module, configured to:

[0170] Determine the working frequency of each employee according to the duration of the image observation information and the time information of each employee in the employee recognition result, wherein the working frequency includes the ratio of the employee's time information to the image observation information;

[0171] Delete employees whose work frequency is lower than the preset threshold in the employee identification results.

[0172] In one embodiment of the present disclosure, the first attribute information includes the type of work clothes;

[0173] The device also includes a deletion module, which is specifically used to:

[0174] Delete employees whose uniform type is not the type to be identified in the employee identification results.

[0175] In one embodiment of the present disclosure, the matching module is used to match at least one area to be identified and at least one employee to be identified according to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, and when obtaining the number of employees and employee identities in each area to be identified, specifically to:

[0176] According to the employee identification results of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, a greedy matching algorithm is used to match at least one area to be identified and at least one employee to be identified to obtain the number and identity of employees in each area to be identified.

[0177] In one embodiment of the present disclosure, the matching module is used to match at least one area to be identified and at least one employee to be identified according to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, and when obtaining the number of employees and employee identities in each area to be identified, specifically to:

[0178] Matching at least one area to be identified and at least one employee to be identified according to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified;

[0179] In response to a successful one-to-one match between at least one to-be-identified area and at least one to-be-identified employee, the number of employees and employee identities in each to-be-identified area according to the matching result;

[0180] In response to the failure of one-to-one matching between at least one area to be identified and at least one employee to be identified, at least one area to be identified and at least one employee to be identified are matched according to time information of at least one employee in the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, so as to obtain the number of employees and employee identities in each area to be identified.

[0181] In one embodiment of the present disclosure, the device further includes an attendance module, which is used to:

[0182] At least one employee to be identified is determined based on the attendance record.

[0183] Regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment of the method in the third aspect, and will not be elaborated here.

[0184] At least one embodiment of the present disclosure further provides a device. Figure 7 , which shows the structure of the device, the device includes a memory and a processor, the memory is used to store computer instructions that can be executed on the processor, and the processor is used to process the order based on the method of the first aspect or the second aspect when executing the computer instructions.

[0185] At least one embodiment of the present disclosure further provides a computer-readable storage medium having a computer program stored thereon, and when the program is executed by a processor, the method of the first aspect or the second aspect is implemented.

[0186] In the present disclosure, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance. The term "plurality" refers to two or more than two, unless otherwise clearly defined.

[0187] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the disclosure disclosed herein. The present disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0188] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. An identity recognition method, characterized in that: The method comprises: Performing recognition processing on the image observation information of at least one area to be recognized, and obtaining a trajectory recognition result in each area to be recognized, wherein the trajectory recognition result includes time information and first attribute information of at least one employee trajectory; Determine a probability value of each employee to be identified appearing in each area to be identified according to the second attribute information of at least one employee to be identified, the third attribute information of the at least one area to be identified, and the trajectory identification result in each area to be identified, wherein the second attribute information includes the identity of the employee; According to the trajectory recognition result in each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, the at least one area to be identified and the at least one employee to be identified are matched to obtain the number of employees and employee identities in each area to be identified.

2. The identity recognition method according to claim 1, characterized in that: The determining, based on the second attribute information of at least one employee to be identified, the third attribute information of at least one area to be identified, and the trajectory recognition result in each area to be identified, the probability value of each employee to be identified appearing in each area to be identified includes: Determining a matching degree between each employee to be identified and each area to be identified based on the second attribute information of the at least one employee to be identified and the third attribute information of the at least one area to be identified; According to the matching degree between each employee to be identified and each area to be identified, and the trajectory identification result in each area to be identified, the probability value of each employee to be identified appearing in each area to be identified is determined.

3. The identity recognition method according to claim 2, characterized in that: Determining the matching degree between each employee to be identified and each area to be identified according to the second attribute information of the at least one employee to be identified and the third attribute information of the at least one area to be identified, including: The matching degree between each employee to be identified and each area to be identified is determined according to the second attribute information of the at least one employee to be identified, the third attribute information of the at least one area to be identified, and the shift information.

4. The identity recognition method according to claim 2, characterized in that: The second attribute information includes employee skills, and the third attribute information includes skills required by the region.

5. The identity recognition method according to any one of claims 2 to 4, characterized in that: The first attribute information includes gender information; The method of determining the probability value of each employee to be identified appearing in each area to be identified according to the matching degree between each employee to be identified and each area to be identified, and the trajectory identification result in each area to be identified, comprises: According to the trajectory recognition results in each area to be identified, the working hours of employees of each gender in each area to be identified are determined accordingly; According to the matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified, the probability value of each employee to be identified appearing in each area to be identified is determined.

6. The identity recognition method according to claim 5, characterized in that: The method of determining the probability value of each employee to be identified appearing in each area to be identified based on the matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified, includes: The matching degree between each employee to be identified and each area to be identified, and the working hours of employees of each gender in each area to be identified, are input into the prediction model to obtain the probability value of each employee to be identified appearing in each area to be identified output by the prediction model.

7. The identity recognition method according to claim 6, characterized in that: The prediction model is obtained through training, and the training includes: Performing recognition processing on historical image observation information of at least one work area to obtain trajectory recognition results in each work area, wherein the historical image observation information of each work area is marked with employee identities; Determining a matching degree between each employee identity and each work area according to the second attribute information corresponding to each employee identity annotated by the historical image observation information of the at least one work area and the third attribute information of the at least one work area; Determine the working hours of employees of each gender in each work area according to the trajectory recognition result in each work area; The prediction model is trained using the matching degree between each employee identity and each work area, the working hours of employees of each gender in each work area of ​​the at least one work area, and the employee identity annotated with the historical image observation information of each work area of ​​the at least one work area.

8. The identity recognition method according to claim 1, characterized in that: According to the trajectory recognition result in each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, matching the at least one area to be identified and the at least one employee to be identified to obtain the number of employees and employee identities in each area to be identified, comprises: For each area to be identified, clustering and merging the at least one employee trajectory according to the time information and the first attribute information of at least one employee trajectory in the trajectory recognition result in the area to be identified, so as to obtain an employee recognition result in the area to be identified, wherein the employee recognition result includes the time information and the first attribute information of at least one employee; According to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, the at least one area to be identified and the at least one employee to be identified are matched to obtain the number and identity of employees in each area to be identified.

9. The identity recognition method according to claim 8, characterized in that: The method further comprises: Determine the working frequency of each employee according to the duration of the image observation information and the time information of each employee in the employee recognition result, wherein the working frequency includes the ratio of the time information of the employee to the image observation information; Delete employees whose working frequency is lower than a preset threshold in the employee identification results.

10. The identity recognition method according to claim 8, characterized in that: The first attribute information includes the type of work clothes; The method further comprises: Delete employees whose uniform type is not the type to be identified in the employee identification results.

11. The identity recognition method according to claim 8, characterized in that: According to the employee recognition result of each to-be-recognized area and the probability value of each to-be-recognized employee appearing in each to-be-recognized area, matching the at least one to-be-recognized area and the at least one to-be-recognized employee to obtain the number of employees and employee identities in each to-be-recognized area includes: According to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, a greedy matching algorithm is used to match the at least one area to be identified and the at least one employee to be identified to obtain the number of employees and employee identities in each area to be identified.

12. The identity recognition method according to claim 8, characterized in that: According to the employee recognition result of each to-be-recognized area and the probability value of each to-be-recognized employee appearing in each to-be-recognized area, matching the at least one to-be-recognized area and the at least one to-be-recognized employee to obtain the number of employees and employee identities in each to-be-recognized area includes: Matching the at least one area to be identified and the at least one employee to be identified according to the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified; In response to a successful one-to-one matching between the at least one area to be identified and the at least one employee to be identified, the number of employees and employee identities in each area to be identified according to the matching result; In response to the at least one area to be identified and the at least one employee to be identified not being successfully matched one-to-one, the at least one area to be identified and the at least one employee to be identified are matched according to the time information of at least one employee in the employee identification result of each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, so as to obtain the number of employees and employee identities in each area to be identified.

13. The identity recognition method according to claim 1, characterized in that: The method further comprises: The at least one employee to be identified is determined according to the attendance record.

14. An identity recognition device, characterized in that: The device comprises: An identification module, used to perform identification processing on the image observation information of at least one area to be identified, and obtain a trajectory identification result in each area to be identified, wherein the trajectory identification result includes time information and first attribute information of at least one employee trajectory; a prediction module, configured to determine a probability value of each to-be-identified employee appearing in each to-be-identified area according to second attribute information of at least one to-be-identified employee, third attribute information of at least one to-be-identified area, and a trajectory recognition result in each to-be-identified area, wherein the second attribute information includes an employee identity; The matching module is used to match the at least one area to be identified and the at least one employee to be identified according to the trajectory recognition result in each area to be identified and the probability value of each employee to be identified appearing in each area to be identified, so as to obtain the number of employees and employee identities in each area to be identified.

15. An electronic device, characterized in that: The device comprises a memory and a processor, wherein the memory is used to store computer instructions executable on the processor, and the processor is used to implement the method according to any one of claims 1 to 13 when executing the computer instructions.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 13 is implemented.

Images