A method, device, electronic device and storage medium for reporting facial images

Abstract

The present invention provides a face image reporting method, device, electronic device and storage medium, which relate to the field of security. The method can perform a first optimization on face images collected in a single continuous tracking of the same face image, and use a feature library to be reported to perform a second optimization on the best target face image of the face in this round of tracking, that is, use the face image to be reported in the feature library to match the target face image, if they match each other, only the one with the best quality is retained, otherwise the target face image is retained; at the same time, the feature library to be reported will temporarily store the face image to be reported, and when it is determined that the temporary storage time of the face image to be reported has expired, it will be removed from the feature library and reported, so as to suppress the multi-shot problem in a short time; in addition, during the reporting process, the present invention will also use a similar reported feature library on the server side to perform another round of matching on the reported face image, so as to further reduce the multi-shot rate and improve the quality of the reported face.

Description

Technical Field

[0001] The present invention relates to the field of security, and in particular to a facial image reporting method, device, electronic equipment and storage medium. Background Art

[0002] In the security field, some places need to perform face recognition and image capture in the monitoring area. However, due to interference from factors such as light and occlusion, the tracking of the target face may be discontinuous, which may easily cause multiple repeated images of the same target. This will have a certain impact on the subsequent face structuring accuracy and will also cause a certain amount of pressure on storage resources.

[0003] Therefore, how to effectively solve the problem of multiple shots of a target to reduce the repetition rate of reported face images is a technical problem that those skilled in the art need to solve. Summary of the invention

[0004] The purpose of the present invention is to provide a facial image reporting method, device, electronic device and storage medium, which can optimize facial images of the same target to reduce the repetition rate and improve the quality of the reported images.

[0005] In order to solve the above technical problems, the present invention provides a method for reporting a facial image, comprising:

[0006] When acquiring facial images collected by continuously tracking the same facial target, calculating the quality value of each of the facial images, and performing feature matching between the target facial image with the highest quality value and the facial images to be reported in the feature library to be reported;

[0007] When it is determined that the target facial image cannot match the facial image to be reported, saving the target facial image to the feature library to be reported and recording the corresponding retention time;

[0008] When the target facial image to be reported that best matches the target facial image is determined, re-recording the retention time of the target facial image to be reported, and retaining the target facial image and the image with the highest quality value among the target facial image to be reported in the feature library to be reported;

[0009] When it is determined that the facial image to be reported whose retention time exceeds the first preset time cannot be matched with the reported facial image in the reported feature library on the server side, the facial image to be reported whose retention time exceeds the first preset time is added to the reported feature library and reported.

[0010] Optionally, calculating the quality value of each of the facial images includes:

[0011] Determine the pitch angle, roll angle and yaw angle corresponding to the face target in the face image, and determine the posture value of the face image using the pitch angle, the roll angle and the yaw angle;

[0012] Determining the clarity value of the face image using the brightness value of the face image;

[0013] Determine the face position value of the face image by using the relative position of the face image in the complete image to which it belongs and the size of the face image;

[0014] The quality value is determined using the posture value, the clarity value and the face position value.

[0015] Optionally, determining the pitch angle, roll angle, and yaw angle corresponding to the face target in the face image includes:

[0016] Position information of facial features of the facial target is located in the facial image, and the pitch angle, the roll angle, and the yaw angle are determined according to the facial feature position information.

[0017] Optionally, performing feature matching on the target facial image with the highest quality value and the facial images to be reported in the feature library to be reported includes:

[0018] Calculating the similarity between the target face image and the face image to be reported to obtain a similarity value, and determining whether the similarity value is greater than a preset threshold;

[0019] If yes, determining that the target face image matches the face image to be reported;

[0020] If not, it is determined that the target facial image does not match the facial image to be reported;

[0021] Accordingly, the step of determining the target face image to be reported that best matches the target face image includes:

[0022] The face image to be reported with the highest similarity value is set as the target face image to be reported.

[0023] Optionally, acquiring the facial images collected by continuously tracking the same facial target includes:

[0024] Acquire continuous image frames, and perform face detection in the image frames to obtain an initial face image;

[0025] The Hungarian algorithm is used to continuously track and match the initial face image to determine the face images in the initial face image that are from the same face target.

[0026] Optionally, the retention time corresponding to the record includes:

[0027] Acquire the acquisition time of the target facial image, and record the retention time of the target facial image according to the acquisition time;

[0028] Correspondingly, the re-recording of the retention time of the target face image to be reported includes:

[0029] The collection time is used to re-record the retention time of the target facial image to be reported.

[0030] Optionally, the step of adding the face image to be reported whose retention time exceeds a first preset time to the reported feature library includes:

[0031] Setting the facial image to be reported whose retention time exceeds the first preset time as a reported facial image, adding the image features of the reported facial image to the reported feature library, and recording the corresponding reporting retention time;

[0032] Accordingly, it also includes:

[0033] When a target reported facial image that best matches the reported facial image is determined, re-recording the reporting retention time of the target reported facial image, and retaining the image features corresponding to the image with the highest quality value between the reported facial image and the target reported facial image in the reported feature library;

[0034] The reported facial images in the reported feature library whose reported retention time exceeds a second preset time are deleted; the second preset time is greater than the first preset time.

[0035] The present invention also provides a facial image reporting device, comprising:

[0036] The quality value calculation and matching module is used to calculate the quality value of each face image when acquiring face images collected by continuously tracking the same face target, and perform feature matching between the target face image with the highest quality value and the face images to be reported in the feature library to be reported;

[0037] A first processing module, configured to save the target facial image to the feature library to be reported and record a corresponding retention time when it is determined that the target facial image cannot match the facial image to be reported;

[0038] A second processing module is used for re-recording the retention time of the target facial image to be reported when determining the target facial image to be reported that best matches the target facial image, and retaining the target facial image and the image with the highest quality value among the target facial image to be reported in the feature library to be reported;

[0039] The reporting module is used to add the face image to be reported whose retention time exceeds the first preset time to the reported feature library and report it when it is determined that the face image to be reported whose retention time exceeds the first preset time cannot match the reported face image in the reported feature library of the server.

[0040] The present invention also provides an electronic device, comprising:

[0041] Memory for storing computer programs;

[0042] A processor is used to implement the above-mentioned facial image reporting method when executing the computer program.

[0043] The present invention also provides a storage medium, in which computer executable instructions are stored. When the computer executable instructions are loaded and executed by a processor, the facial image reporting method as described above is implemented.

[0044] The present invention provides a method for reporting facial images, comprising: when facial images collected by continuously tracking the same facial target are obtained, the quality value of each facial image is calculated, and the target facial image with the highest quality value is feature matched with the facial image to be reported in a feature library to be reported; when it is determined that the target facial image cannot be matched with the facial image to be reported, the target facial image is saved in the feature library to be reported and the corresponding retention time is recorded; when the target facial image to be reported that best matches the target facial image is determined, the retention time of the target facial image to be reported is re-recorded, and the target facial image and the image with the highest quality value among the target facial images to be reported are retained in the feature library to be reported; when it is determined that the facial image to be reported whose retention time exceeds a first preset time cannot be matched with the reported facial image in the reported feature library of the server side, the facial image to be reported whose retention time exceeds the first preset time is added to the reported feature library and reported.

[0045] It can be seen that when the present invention obtains the face images collected by continuously tracking the same face target, the quality value of each face image can be calculated, and the target face image with the highest quality value can be selected therefrom, that is, the face images collected when continuously tracking the same face target can be screened for the first round according to the quality value; then, the present invention will use the target face image to perform feature matching with the face images to be reported in the feature library to be reported to determine whether the target face image has been saved in the feature library to be reported. If the match is not successful, the target face image will be saved in the feature library to be reported, and the corresponding retention time will be recorded; if the match is successful, the retention time corresponding to the target face image to be reported that best matches the target face image will be re-recorded, and only the image with the highest quality value will be retained between the target face image and the target face image to be reported, and then the retention time exceeding the first predicted value will be reported. The face image to be reported is set to a time. In this way, for the multi-shot images collected by multiple discontinuous tracking of the same target in a short period of time, the feature library to be reported can be used for repeated filtering, and the same target only saves one high-quality face image in the feature library to be reported. Not only can this high-quality face image be used for repeated filtering to solve the multi-shot problem caused by discontinuous tracking, but also the burden on the storage device can be effectively reduced; finally, before reporting the face image to be reported whose retention time exceeds the first preset time, the present invention will also match it with the reported face image in the reported feature library on the server side to determine whether there is a face image of the same target that has been reported before this face image to be reported. If it cannot be matched, it will be reported, which can effectively suppress the multi-shot problem caused by the long-term stay of the personnel and further reduce the multi-shot rate. The present invention also provides a face image reporting device, an electronic device and a storage medium, which have the above-mentioned beneficial effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] 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 embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0047] Figure 1 A flowchart of a facial image reporting method provided by an embodiment of the present invention;

[0048] Figure 2 A flowchart of another facial image reporting method provided by an embodiment of the present invention;

[0049] Figure 3 This is a structural block diagram of a facial image reporting device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] In the field of security, some places need to perform face recognition and image capture in the monitoring area. However, due to interference from factors such as light and occlusion, the tracking of the target face may be discontinuous, which may easily cause multiple repeated images of the same target, which will have a certain impact on the accuracy of subsequent face structuring and will also cause certain pressure on storage resources. In view of this, the present invention provides a face image reporting method that can optimize the face images of the same target to reduce the repetition rate and improve the image quality. Please refer to Figure 1 , Figure 1 A flowchart of a facial image reporting method provided by an embodiment of the present invention, the method may include:

[0052] S101. When facial images collected by continuously tracking the same facial target are obtained, the quality value of each facial image is calculated, and the target facial image with the highest quality value is feature matched with the facial images to be reported in the feature library to be reported.

[0053] In the related art, in order to collect the facial images corresponding to the same facial target in real time, the facial target is usually tracked to obtain the continuous facial images corresponding to the target. When the present application obtains these facial images, the quality values ​​corresponding to these facial images are first calculated to evaluate and select these facial images from multiple angles, so as to initially select the facial images collected by the single continuous tracking of the same facial target according to the quality value. In other words, the facial target can obtain a single facial image in the complete single continuous tracking process from entering the tracked state to the interruption of the tracking state. However, due to the influence of factors such as light and obstructions, the continuous tracking of the facial target may be temporarily interrupted, and then the same facial target may correspond to multiple continuous trackings and multiple preferred facial images. If all these preferred images are reported, it is easy to have a certain impact on the accuracy of the subsequent facial structure and will cause a certain pressure on storage resources. For this reason, the embodiment of the present invention also additionally sets up a feature library to be reported, which can temporarily store facial images to be uploaded, and then whenever a preferred facial image is obtained, it can be feature matched with the facial images to be reported in the library to determine whether the facial image corresponding to the same target has been collected, and then the facial image of the same facial target can be optimized twice to achieve the purpose of reducing the repetition rate of facial images. Of course, in order to further improve the matching efficiency, in addition to storing the facial image itself in the feature library to be reported, the image features corresponding to the facial image can also be stored therein. The embodiment of the present invention does not limit the specific image features, and reference may be made to related technologies of image recognition.

[0054] Furthermore, it should be noted that the embodiments of the present invention do not limit the specific method of face tracking. For example, the Hungarian algorithm can be used for tracking. Of course, it is understandable that in order to perform continuous face tracking, it is first necessary to perform face detection in continuous image frames to obtain an initial face image, and then the Hungarian algorithm can be used to perform feature matching on the corresponding initial face images in adjacent image frames to determine the face image from the same face target, that is, to achieve continuous tracking of the target. The embodiments of the present invention do not limit the specific method of face detection, and reference can be made to related technologies of image recognition and face recognition.

[0055] In one possible case, obtaining face images collected by continuously tracking the same face target includes:

[0056] Step 11: Acquire continuous image frames, and perform face detection in the image frames to obtain an initial face image;

[0057] Step 12: Use the Hungarian algorithm to continuously track and match the initial face image to determine the face images from the same face target in the initial face image.

[0058] Furthermore, the embodiment of the present invention does not limit the calculation method of the above-mentioned quality value. For example, the orientation posture of the face target in the face image can be determined (such as determining the pitch angle, roll angle and yaw angle corresponding to the face in the face image), and the quality value of the face image can be calculated based on the orientation posture; the quality value of the face image can also be determined based on the brightness of the face image; the quality value of the face image can also be determined based on the relative position and size of the face image in the complete image. Of course, the above-mentioned information can also be used to determine the quality value of the face image at the same time, which can be selected according to actual application requirements. In the embodiment of the present invention, in order to perform an all-round evaluation of the face image, the quality value can be calculated together using the posture value calculated by the posture information of the face target in the face image, the clarity value calculated by the brightness value of the face image, and the face position value determined by the relative position of the face image in the complete image and the size of the face image.

[0059] In one possible case, calculating the quality value of each face image may include:

[0060] Step 21: determining the pitch angle, roll angle and yaw angle corresponding to the face target in the face image, and determining the posture value of the face image using the pitch angle, roll angle and yaw angle;

[0061] Step 22: Determine the clarity value of the face image using the brightness value of the face image;

[0062] Step 23: Determine the face position value of the face image by using the relative position of the face image in the complete image and the size of the face image;

[0063] Step 24: Determine the quality value using the pose value, clarity value, and face position value.

[0064] It should be noted that the embodiments of the present invention do not limit the specific calculation method of the above-mentioned posture value, clarity value, face position value and quality value. For example, weighted calculation can be used for calculation, which can be selected according to actual application requirements.

[0065] Furthermore, it should be noted that the embodiments of the present invention do not limit how to determine the pitch angle, roll angle and yaw angle corresponding to the facial target in the facial image. For example, the facial features of the facial target can be located first to determine the position information of the facial features, and then the pitch angle, roll angle and yaw angle are determined based on the position information.

[0066] In one possible case, determining the pitch angle, roll angle, and yaw angle corresponding to the face target in the face image may include:

[0067] Step 31: Locate the facial features of the face target in the face image, and determine the pitch angle, roll angle, and yaw angle based on the facial features position information.

[0068] Furthermore, it should be noted that the embodiments of the present invention do not limit the specific feature matching method between the target facial image and the facial image to be reported. For example, the similarity between the target facial image and the facial image to be reported can be calculated, and whether the target facial image matches the facial image to be reported can be determined based on whether the similarity exceeds a preset threshold. It should be noted that the embodiments of the present invention do not limit the specific method for calculating the similarity value. For example, the calculation can be performed based on the image features of the target facial image and the facial image to be reported. For details, reference can be made to the relevant technology of graphic recognition.

[0069] In one possible case, feature matching of the target face image with the highest quality value with the face images to be reported in the feature library to be reported may include:

[0070] Step 41: Calculate the similarity between the target face image and the face image to be reported to obtain a similarity value, and determine whether the similarity value is greater than a preset threshold; if so, proceed to step 42; if not, proceed to step 43;

[0071] Step 42: Determine whether the target face image matches the face image to be reported;

[0072] Step 43: Determine whether the target facial image does not match the facial image to be reported.

[0073] S102: When it is determined that the target facial image cannot match the facial image to be reported, the target facial image is saved in a feature library to be reported and a corresponding retention time is recorded.

[0074] When it is determined that the target facial image does not match any of the facial images to be reported, it can be determined that the facial target corresponding to the target facial image appears in the monitoring area for the first time, that is, its facial image needs to be reported, so the target facial image can be saved in the feature library to be reported for temporary storage. It should be noted that the retention time that appears in step S102 is the time that the target facial image has been saved in the feature library to be reported, that is, it needs to be recorded in real time. The embodiment of the present invention does not limit which moment is used as the starting moment for the above-mentioned retention time. For example, the moment when the target facial image is temporarily stored in the feature library to be reported can be used as the starting moment, or the moment when the target facial image is captured by the image acquisition device can be used as the starting moment. In the embodiment of the present invention, in order to improve the real-time nature of the report, the capture moment of the target facial image can be used as the above-mentioned starting moment.

[0075] In one possible scenario, the corresponding retention time of the record may include:

[0076] Step 51: Obtain the acquisition time of the target face image, and record the retention time of the target face image according to the acquisition time.

[0077] S103. When the target facial image to be reported that best matches the target facial image is determined, the retention time of the target facial image to be reported is re-recorded, and the target facial image and the image with the highest quality value among the target facial images to be reported are retained in the feature library to be reported.

[0078] In an embodiment of the present invention, after determining that the target face image and the face image to be reported match, the target face image and the target face image to be reported that best matches the target face image are selected to retain a higher quality image in the feature library to be reported. In other words, only one best face image is saved in the feature library to be reported for the same face target. The advantage of such a setting is that this best face image can contain enough features, not only can it be matched with other face images to determine whether they are from the same face target, but more importantly, only a small number of images need to be saved in the feature library to be reported to achieve the above purpose, which can effectively reduce the pressure on storage space caused by face image deduplication. Further, it can be understood that since the target face image and the target face image to be reported both correspond to the same face target, that is, the face image corresponding to the face target is obtained again, it is necessary to re-record the retention time of the target face image to be reported in order to continue to temporarily store the face image corresponding to the target. Similarly, the retention time can also be re-recorded based on the acquisition time of the target face image.

[0079] In a possible case, re-recording the retention time of the target facial image to be reported may include:

[0080] Step 61: Use the acquisition time to re-record the retention time of the target facial image to be reported.

[0081] Further, it is understandable that when the target facial image and each facial image to be reported are matched for features using similarity values, after determining that the match is successful, the successfully matched image can be set as the above-mentioned target facial image to be reported. However, since the target facial image needs to be compared with the facial images in the library one by one, there may be multiple facial images to be reported whose similarity values ​​with the target facial image are greater than a preset threshold. In this case, the facial image to be reported with the highest similarity value can generally be selected as the target facial image to be reported.

[0082] In one possible scenario, determining the target face image to be reported that best matches the target face image may include:

[0083] Step 71: setting the to-be-reported facial image with the highest similarity value as the target to-be-reported facial image.

[0084] S104. When it is determined that the facial image to be reported whose retention time exceeds the first preset time cannot be matched with the reported facial image in the reported feature library on the server side, the facial image to be reported whose retention time exceeds the first preset time is added to the reported feature library and reported.

[0085] It should be noted that the embodiment of the present invention does not limit the specific value of the first preset time, which can be set according to actual application requirements, for example, it can be 10 seconds.

[0086] After determining that the retention time exceeds the first preset time, it can be determined that the face target has left the monitoring area. At this time, the corresponding face image to be reported can be removed from the feature library to be reported and reported. However, in actual application, there may be situations where people stay for a long time in some scenarios. For example, in subway scenarios, there may be situations where staff stay at their workstations for a long time. This situation will also lead to serious multi-shot phenomenon. In view of this, the embodiment of the present invention also sets up an additional reported feature library, which stores the reported face images that have been reported. In this way, before reporting the face image to be reported whose retention time exceeds the first preset time, it can be matched with the reported face image in the reported feature library. If the match is successful, it means that other face images of the same target have been uploaded before this face image; if the match is unsuccessful, it means that other similar face images have not been uploaded before this face image. At this time, the face image to be reported can be reported and saved in the reported feature library. In this way, the multi-shot problem caused by the long stay of personnel can be further suppressed. It should be noted that the embodiment of the present invention does not limit the specific setting form of the reported feature library, which can be similar to the feature library to be reported, that is, it is used to temporarily store reported face images. When it is found that the reporting retention time corresponding to a reported face image exceeds the preset time, it is removed; at the same time, the same face target only stores one best reported face image in the reported feature library, that is, when the target reported face image that best matches the face image to be reported is determined, only the one with the highest quality value of the two images is retained, and the reporting retention time corresponding to the target reported face image is re-recorded. In this way, the storage space occupied can be minimized under the premise of suppressing multiple shots. Of course, in order to further reduce the storage space occupied, the reported feature library can even only save the image features of the reported face image without saving the image itself; further, when the target reported face image that best matches the face image to be reported is determined, only the image features corresponding to the one with the highest quality value of the two images are retained. Furthermore, it can be understood that the temporary storage time corresponding to the reported feature library should be greater than the temporary storage time corresponding to the to-be-reported feature library (ie, the first preset time).

[0087] Furthermore, it should be pointed out that the feature library to be reported can be set at the image acquisition end, while the reported feature library is set at the server end. The image acquisition end needs to send the reported face image to the server end so that the server end can use the reported feature library to suppress the capture of this image, and only capture the reported face image that cannot match any reported face image in the reported feature library.

[0088] Based on the above embodiments, when acquiring facial images collected by continuously tracking the same facial target, the present invention can calculate the quality value of each of the facial images and select the target facial image with the highest quality value therefrom, that is, it can perform a first round of screening on the facial images collected when continuously tracking the same facial target according to the quality value; then, the present invention will use the target facial image to perform feature matching with the facial images to be reported in the feature library to be reported to determine whether the target facial image has been saved in the feature library to be reported. If the match is not successful, the target facial image will be saved in the feature library to be reported, and the corresponding retention time will be recorded; if the match is successful, the retention time corresponding to the target facial image to be reported that best matches the target facial image will be re-recorded, and only the image with the highest quality value will be retained between the target facial image and the target facial image to be reported, and then the retention time exceeding the target facial image will be reported. The facial image to be reported within the first preset time, in this way, for the multi-shot images collected by multiple discontinuous tracking of the same target in a short period of time, the feature library to be reported can be used for repeated filtering, and only one high-quality facial image of the same target is stored in the feature library to be reported. Not only can this high-quality facial image be used for repeated filtering to solve the multi-shot problem caused by discontinuous tracking, but also the burden on the storage device can be effectively reduced; finally, before reporting the facial image to be reported whose retention time exceeds the first preset time, the present invention will also match it with the reported facial images in the reported feature library on the server side to determine whether there is a facial image of the same target that has been reported before this facial image to be reported. If no match can be made, it will be reported, which can effectively suppress the multi-shot problem caused by long-term stay of personnel and further reduce the multi-shot rate.

[0089] Based on the above embodiment, the specific setting method of the reported feature library is described in detail below. Figure 2 , Figure 2 A flowchart of another facial image reporting method provided by an embodiment of the present invention. In one possible scenario, the facial image reporting method may include:

[0090] S201. When facial images collected by continuously tracking the same facial target are obtained, the quality value of each facial image is calculated, and the target facial image with the highest quality value is matched with the facial images to be reported in the feature library to be reported for feature matching.

[0091] S202: When it is determined that the target facial image cannot be matched with the facial image to be reported, the target facial image is saved in the feature library to be reported and the corresponding retention time is recorded.

[0092] S203. When the target facial image to be reported that best matches the target facial image is determined, the retention time of the target facial image to be reported is re-recorded, and the target facial image and the image with the highest quality value among the target facial images to be reported are retained in the feature library to be reported.

[0093] It should be noted that the relevant descriptions of S201 to S203 are consistent with those in the above embodiment. Please refer to the descriptions in the above embodiment and no further details will be given here.

[0094] S204: setting the facial image to be reported whose retention time exceeds the first preset time as the facial image to be reported, and matching the facial image to be reported with the reported facial images in the reported feature library on the server side for feature matching.

[0095] S205. When it is determined that the reported facial image cannot match the reported facial images in the reported feature library, the image features of the reported facial image are added to the reported feature library, and the corresponding reporting retention time is recorded.

[0096] In the embodiment of the present invention, a reported feature library is further set on the server side on the basis of setting a feature library to be reported. The feature library is used to store reported facial images, and to suppress snapshots of reported facial images sent by the image acquisition end, and is set with a longer temporary storage time. In this way, when the server receives a reported facial image, it can use the reported facial images in the reported feature library to perform feature matching with it to determine whether the facial image to be processed has been reported; if it has been reported, it will be optimized with the best matching target reported facial image to retain the image features corresponding to the facial image with higher quality in the reported feature library; if it has not been reported, the reporting operation will be performed, and the image features of the reported facial image will be saved in the reported feature library.

[0097] It should be noted that the reporting retention time is the same as the retention time in the above embodiment, and reference may be made to the relevant description in the above embodiment.

[0098] S206. When a target reported facial image that best matches the reported facial image is determined, the reporting retention time of the target reported facial image is re-recorded, and the image features corresponding to the image with the highest quality value between the reported facial image and the target reported facial image are retained in the reported feature library.

[0099] S207. Delete the reported facial images in the reported feature database whose reported retention time exceeds a second preset time; the second preset time is greater than the first preset time.

[0100] It should be noted that the temporary storage time of the reported feature library is relatively long, that is, the second preset time should be greater than the first preset time, so that the problem of multiple shots caused by long-term stay of personnel can be suppressed. It should be noted that the embodiment of the present invention does not limit the specific value of the second preset time, and it can be set according to actual application requirements.

[0101] Based on the above embodiments, the embodiments of the present invention further adopt a reported feature library to suppress snapshots of reported facial images to solve the problem of multiple shots caused by long-term stay of personnel and reduce the storage pressure on the server side.

[0102] The following is an introduction to the facial image reporting device, electronic device and storage medium provided in the embodiments of the present invention. The facial image reporting device, electronic device and storage medium described below can be referenced to the facial image reporting method described above.

[0103] Please refer to Figure 3 , Figure 3 A structural block diagram of a facial image reporting device provided by an embodiment of the present invention, the device may include:

[0104] The quality value calculation and matching module 301 is used to calculate the quality value of each face image when acquiring face images collected by continuously tracking the same face target, and perform feature matching between the target face image with the highest quality value and the face images to be reported in the feature library to be reported;

[0105] The first processing module 302 is used to save the target facial image to the feature library to be reported and record the corresponding retention time when it is determined that the target facial image cannot match the facial image to be reported;

[0106] The second processing module 303 is used for re-recording the retention time of the target facial image to be reported when determining the target facial image to be reported that best matches the target facial image, and retaining the target facial image and the image with the highest quality value in the feature library to be reported;

[0107] The reporting module 304 is used to add the facial image to be reported whose retention time exceeds the first preset time to the reported feature library and report it when it is determined that the facial image to be reported whose retention time exceeds the first preset time cannot be matched with the reported facial image in the reported feature library on the server side.

[0108] Optionally, the quality value calculation and matching module 301 may include:

[0109] The posture value calculation submodule is used to determine the pitch angle, roll angle and yaw angle corresponding to the face target in the face image, and use the pitch angle, roll angle and yaw angle to determine the posture value of the face image;

[0110] A clarity value calculation submodule, used to determine the clarity value of the face image using the brightness value of the face image;

[0111] A face position value calculation submodule is used to determine the face position value of the face image by using the relative position of the face image in the complete image to which it belongs and the size of the face image;

[0112] The quality value calculation submodule is used to determine the quality value using the posture value, the clarity value and the face position value.

[0113] Optionally, the posture value calculation submodule may include:

[0114] The posture determination unit is used to locate the facial features of the face target in the face image, and determine the pitch angle, roll angle and yaw angle based on the facial features position information.

[0115] Optionally, the quality value calculation and matching module 301 may include:

[0116] The similarity value calculation submodule is used to calculate the similarity between the target face image and the face image to be reported to obtain a similarity value, and determine whether the similarity value is greater than a preset threshold;

[0117] A first determination module, configured to determine that if yes, the target face image matches the face image to be reported;

[0118] A second determination module is used to determine that the target face image does not match the face image to be reported if no;

[0119] Accordingly, the second processing module 303 may include:

[0120] The setting submodule is used to set the face image to be reported with the highest similarity value as the target face image to be reported.

[0121] Optionally, the quality value calculation and matching module 301 includes:

[0122] The face detection submodule is used to obtain continuous image frames and perform face detection in the image frames to obtain an initial face image;

[0123] The tracking submodule is used to continuously track and match the initial face image using the Hungarian algorithm to determine the face images from the same face target in the initial face image.

[0124] Optionally, the first processing module 302 may include:

[0125] The first retention time recording submodule is used to obtain the acquisition time of the target face image and record the retention time of the target face image according to the acquisition time;

[0126] Accordingly, the second processing module 303 may include:

[0127] The second retention time recording submodule is used to re-record the retention time of the target face image to be reported using the collection time.

[0128] Optionally, the reporting module 304 is specifically configured to:

[0129] When it is determined that the face image to be reported whose retention time exceeds the first preset time cannot be matched with the reported face image in the reported feature library of the server, the face image to be reported whose retention time exceeds the first preset time is set as the reported face image, the image features of the reported face image are added to the reported feature library, and the corresponding reporting retention time is recorded;

[0130] Accordingly, the device may further include:

[0131] An updating module, for re-recording the reporting retention time of the target reported facial image when determining the target reported facial image that best matches the reported facial image, and retaining the image features corresponding to the image with the highest quality value between the reported facial image and the target reported facial image in the reported feature library;

[0132] The deletion module is used to delete the reported face images in the reported feature library whose reported retention time exceeds the second preset time; the second preset time is greater than the first preset time.

[0133] An embodiment of the present invention further provides an electronic device, including:

[0134] Memory for storing computer programs;

[0135] A processor is used to implement the steps of the above-mentioned facial image reporting method when executing a computer program.

[0136] Since the embodiments of the electronic device part correspond to the embodiments of the facial image reporting method part, please refer to the description of the embodiments of the facial image reporting method part for the embodiments of the electronic device part, which will not be repeated here.

[0137] An embodiment of the present invention further provides a storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the facial image reporting method of any of the above embodiments are implemented.

[0138] Since the embodiments of the storage medium part correspond to the embodiments of the face image reporting method part, please refer to the description of the embodiments of the face image reporting method part for the embodiments of the storage medium part, which will not be repeated here.

[0139] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0140] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0141] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0142] The above is a detailed introduction to a facial image reporting method, device, electronic device and storage medium provided by the present invention. This article uses specific examples to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present invention, several improvements and modifications can be made to the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

Claims

1. A method for reporting a facial image, characterized in that: include: When acquiring facial images collected by continuously tracking the same facial target, calculating the quality value of each of the facial images, and performing feature matching between the target facial image with the highest quality value and the facial images to be reported in the feature library to be reported; When it is determined that the target facial image cannot match the facial image to be reported, saving the target facial image to the feature library to be reported and recording the corresponding retention time; When the target facial image to be reported that best matches the target facial image is determined, re-recording the retention time of the target facial image to be reported, and retaining the target facial image and the image with the highest quality value among the target facial image to be reported in the feature library to be reported; When it is determined that the to-be-reported facial image whose retention time exceeds the first preset time cannot be matched with the reported facial image in the reported feature library of the server, the to-be-reported facial image whose retention time exceeds the first preset time is added to the reported feature library and reported; The step of adding the face image to be reported whose retention time exceeds the first preset time to the reported feature library includes: Setting the facial image to be reported whose retention time exceeds the first preset time as a reported facial image, adding the image features of the reported facial image to the reported feature library, and recording the corresponding reporting retention time; The reported facial images in the reported feature library whose reported retention time exceeds a second preset time are deleted; the second preset time is greater than the first preset time.

2. The facial image reporting method according to claim 1, characterized in that: The calculating the quality value of each of the face images comprises: Determine the pitch angle, roll angle and yaw angle corresponding to the face target in the face image, and determine the posture value of the face image using the pitch angle, the roll angle and the yaw angle; Determining the clarity value of the face image using the brightness value of the face image; Determine the face position value of the face image by using the relative position of the face image in the complete image to which it belongs and the size of the face image; The quality value is determined using the posture value, the clarity value and the face position value.

3. The facial image reporting method according to claim 2, characterized in that: The determining the pitch angle, roll angle and yaw angle corresponding to the face target in the face image includes: Position information of facial features of the facial target is located in the facial image, and the pitch angle, the roll angle, and the yaw angle are determined according to the facial feature position information.

4. The facial image reporting method according to claim 1, characterized in that: The feature matching of the target face image with the highest quality value with the face images to be reported in the feature library to be reported includes: Calculating the similarity between the target face image and the face image to be reported to obtain a similarity value, and determining whether the similarity value is greater than a preset threshold; If yes, determining that the target face image matches the face image to be reported; If not, it is determined that the target facial image does not match the facial image to be reported; Accordingly, the step of determining the target face image to be reported that best matches the target face image includes: The face image to be reported with the highest similarity value is set as the target face image to be reported.

5. The facial image reporting method according to claim 1, characterized in that: The acquiring of the face images collected by continuously tracking the same face target includes: Acquire continuous image frames, and perform face detection in the image frames to obtain an initial face image; The Hungarian algorithm is used to continuously track and match the initial face image to determine the face images in the initial face image that are from the same face target.

6. The facial image reporting method according to claim 1, characterized in that: The retention time corresponding to the records includes: Acquire the acquisition time of the target facial image, and record the retention time of the target facial image according to the acquisition time; Correspondingly, the re-recording of the retention time of the target face image to be reported includes: The collection time is used to re-record the retention time of the target facial image to be reported.

7. The facial image reporting method according to any one of claims 1 to 6, characterized in that: Also includes: When a target reported facial image that best matches the reported facial image is determined, the reporting retention time of the target reported facial image is re-recorded, and the image features corresponding to the image with the highest quality value between the reported facial image and the target reported facial image are retained in the reported feature library.

8. A facial image reporting device, characterized in that: include: The quality value calculation and matching module is used to calculate the quality value of each face image when acquiring face images collected by continuously tracking the same face target, and perform feature matching between the target face image with the highest quality value and the face images to be reported in the feature library to be reported; A first processing module, configured to save the target facial image to the feature library to be reported and record a corresponding retention time when it is determined that the target facial image cannot match the facial image to be reported; A second processing module is used for re-recording the retention time of the target facial image to be reported when determining the target facial image to be reported that best matches the target facial image, and retaining the target facial image and the image with the highest quality value among the target facial image to be reported in the feature library to be reported; A reporting module, configured to add the facial image to be reported whose retention time exceeds the first preset time to the reported feature library and report it when it is determined that the facial image to be reported whose retention time exceeds the first preset time cannot be matched with the reported facial image in the reported feature library of the server; The reporting module is specifically used for: Setting the facial image to be reported whose retention time exceeds the first preset time as a reported facial image, adding the image features of the reported facial image to the reported feature library, and recording the corresponding reporting retention time; A deletion module is used to delete the reported facial images in the reported feature library whose reported retention time exceeds a second preset time; the second preset time is greater than the first preset time.

9. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the facial image reporting method as described in any one of claims 1 to 7 when executing the computer program.

10. A storage medium, characterized in that: The storage medium stores computer executable instructions, and when the computer executable instructions are loaded and executed by the processor, the facial image reporting method according to any one of claims 1 to 7 is implemented.

Images