Image processing method, device, apparatus and storage medium

Abstract

The present disclosure relates to an image processing method and device, equipment and storage medium, and relates to the technical field of computer, which can improve the accuracy of face detection. The image processing method comprises: acquiring a to-be-detected image; inputting the to-be-detected image into a face detection model to obtain a first confidence and feature information of the to-be-detected image; the first confidence represents the probability that a face exists in the to-be-detected image; in the case that the first confidence is greater than a first confidence threshold and less than a second confidence threshold, inputting the feature information of the to-be-detected image into a local detection model to obtain local feature information and a second confidence; the second confidence represents the probability that a local face feature exists in the feature information input into the local detection model; and determining whether a face exists in the to-be-detected image according to a confidence set comprising the first confidence and the second confidence.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of computer, and particularly relates to an image processing method and device, equipment and storage medium. BACKGROUND

[0002] Face detection refers to searching a given image by using a certain strategy to determine whether the image contains a face, and returning the position, size and pose of the face if the image contains a face. Face detection is the basis of many computer vision (CV) tasks, and the accuracy of face detection determines the final effect of subsequent tasks (such as face recognition, face attribute recognition and expression recognition) to a great extent.

[0003] The existing face detection method usually has the problems of missed detection and false detection to a certain extent. Therefore, how to accurately recognize the face in the image is a technical problem to be solved at present. SUMMARY

[0004] The present disclosure provides an image processing method, device, equipment and storage medium, which can accurately recognize the face in the image.

[0005] The technical solutions of the embodiments of the present disclosure are as follows:

[0006] According to a first aspect of the embodiments of the present disclosure, an image processing method is provided, which can be applied to an electronic device. The method can include:

[0007] obtaining a to-be-detected image;

[0008] inputting the to-be-detected image into a face detection model to obtain a first confidence and feature information of the to-be-detected image; the first confidence represents a probability that a face exists in the to-be-detected image;

[0009] in a case where the first confidence is greater than a first confidence threshold and less than a second confidence threshold, inputting the feature information of the to-be-detected image into a local detection model to obtain local feature information and a second confidence; the second confidence represents a probability that a local feature of a face exists in the feature information input into the local detection model;

[0010] determining whether a face exists in the to-be-detected image according to a confidence set including the first confidence and the second confidence.

[0011] Optionally, the local detection model includes at least one of a left eye detection model, a right eye detection model and a mouth detection model.

[0012] In a case where the local detection model includes a left eye detection model, the second confidence includes a left eye confidence output by the left eye detection model; the left eye confidence is used to represent a probability that a left eye feature exists in the left eye local feature information input into the left eye detection model;

[0013] In a case where the local detection model includes a right eye detection model, the second confidence includes a right eye confidence output by the right eye detection model; the right eye confidence is used to represent a probability that a right eye feature exists in the right eye local feature information input into the right eye detection model;

[0014] In a case where the local detection model includes a mouth detection model, the second confidence includes a mouth confidence output by the mouth detection model; the mouth confidence is used to represent a probability that a mouth feature exists in the mouth local feature information input into the mouth detection model.

[0015] Optionally, the face detection model includes a face detection sub-model used to determine a probability that a face exists in the to-be-detected image, a face orientation model used to correct a face orientation, and a face key point detection model used to detect face key points; the feature information of the to-be-detected image further includes face key point position information in the to-be-detected image.

[0016] The to-be-detected image is input into the face detection model to obtain the first confidence and the feature information of the to-be-detected image, including:

[0017] The to-be-detected image is input into the face detection sub-model to obtain face boundary information and the first confidence;

[0018] The face boundary information is input into the face orientation model to obtain a corrected to-be-detected image;

[0019] The corrected to-be-detected image is input into the face key point detection model to obtain a third confidence and face key point position information; the third confidence represents a probability that a face exists in the corrected to-be-detected image.

[0020] Optionally, the local feature information further includes local key point position information and a local key point occlusion state; the occlusion state is unoccluded or occluded;

[0021] The feature information of the to-be-detected image is input into the local detection model to obtain the local feature information and the second confidence, including:

[0022] When the face key point position information includes left eye position information, the left eye position information is input into the left eye detection model to obtain a left eye confidence, left eye key point position information in the left eye local feature information, and a left eye key point occlusion state;

[0023] and / or, when the facial key point position information comprises right eye position information, inputting the right eye position information into a right eye detection model to obtain right eye confidence, right eye key point position information in right eye local feature information, and a right eye key point occlusion state;

[0024] and / or, when the facial key point position information comprises mouth position information, inputting the mouth position information into a mouth detection model to obtain mouth confidence, mouth key point position information in mouth local feature information, and a mouth key point occlusion state.

[0025] Optionally, the confidence set further comprises a third confidence;

[0026] According to the confidence set comprising the first confidence and the second confidence, determining whether a face exists in the to-be-detected image, comprising:

[0027] According to the first confidence, the second confidence, the third confidence, and a preset algorithm, determining a target value;

[0028] When the target value is greater than or equal to a preset threshold, determining that the face exists in the to-be-detected image;

[0029] Or, when the target value is less than the preset threshold, determining that the face does not exist in the to-be-detected image.

[0030] Optionally, the image processing method further comprises:

[0031] When the first confidence is greater than or equal to a second confidence threshold, determining that the face exists in the to-be-detected image, and outputting the feature information of the to-be-detected image;

[0032] Or, when the first confidence is less than or equal to a first confidence threshold, determining that the face does not exist in the to-be-detected image.

[0033] According to a second aspect of the embodiments of the present disclosure, an image processing apparatus is provided, and the method can be applied to an electronic device. The apparatus can comprise an acquisition unit and a processing unit;

[0034] The acquisition unit is configured to acquire a to-be-detected image;

[0035] The processing unit is configured to input the to-be-detected image into a face detection model to obtain a first confidence and feature information of the to-be-detected image; the first confidence represents a probability that a face exists in the to-be-detected image;

[0036] The processing unit is further configured to, in a case where the first confidence is greater than a first confidence threshold and less than a second confidence threshold, input the feature information of the to-be-detected image into a local detection model to obtain local feature information and a second confidence; the second confidence represents a probability that a local feature of a face exists in the feature information input into the local detection model;

[0037] The processing unit is further configured to determine whether the face exists in the to-be-detected image according to the confidence set including the first confidence and the second confidence.

[0038] Optionally, the local detection model includes at least one of a left eye detection model, a right eye detection model, and a mouth detection model.

[0039] In a case where the local detection model includes the left eye detection model, the second confidence includes a left eye confidence output by the left eye detection model; the left eye confidence is used to represent a probability that the left eye feature exists in the left eye local feature information input into the left eye detection model.

[0040] In a case where the local detection model includes the right eye detection model, the second confidence includes a right eye confidence output by the right eye detection model; the right eye confidence is used to represent a probability that the right eye feature exists in the right eye local feature information input into the right eye detection model.

[0041] In a case where the local detection model includes the mouth detection model, the second confidence includes a mouth confidence output by the mouth detection model; the mouth confidence is used to represent a probability that the mouth feature exists in the mouth local feature information input into the mouth detection model.

[0042] Optionally, the face detection model includes a face detection sub-model used to determine a probability that the face exists in the to-be-detected image, a face orientation model used to correct the face orientation, and a face key point detection model used to detect the face key points; the feature information of the to-be-detected image further includes face key point position information in the to-be-detected image.

[0043] The processing unit is specifically configured to:

[0044] input the to-be-detected image into the face detection sub-model to obtain face boundary information and the first confidence;

[0045] input the face boundary information into the face orientation model to obtain a corrected to-be-detected image;

[0046] input the corrected to-be-detected image into the face key point detection model to obtain third confidence and face key point position information; the third confidence represents a probability that the face exists in the corrected to-be-detected image.

[0047] Optionally, the local feature information further includes local key point position information and an occlusion state of the local key point; the occlusion state is unoccluded or occluded.

[0048] The processing unit is specifically configured to:

[0049] When the face key point position information comprises left eye position information, the left eye position information is input into the left eye detection model to obtain left eye confidence, left eye key point position information in left eye local feature information, and a left eye key point occlusion state.

[0050] When the face key point position information comprises right eye position information, the right eye position information is input into the right eye detection model to obtain right eye confidence, right eye key point position information in right eye local feature information, and a right eye key point occlusion state.

[0051] When the face key point position information comprises mouth position information, the mouth position information is input into the mouth detection model to obtain mouth confidence, mouth key point position information in mouth local feature information, and a mouth key point occlusion state.

[0052] Optionally, the confidence set further comprises a third confidence.

[0053] The processing unit is specifically configured to:

[0054] determine a target value according to the first confidence, the second confidence, the third confidence, and a preset algorithm.

[0055] When the target value is greater than or equal to a preset threshold, it is determined that the face exists in the to-be-detected image.

[0056] Or, when the target value is less than the preset threshold, it is determined that the face does not exist in the to-be-detected image.

[0057] Optionally, the processing unit is further configured to:

[0058] When the first confidence is greater than or equal to a second confidence threshold, it is determined that the face exists in the to-be-detected image, and the feature information of the to-be-detected image is output.

[0059] Or, when the first confidence is less than or equal to a first confidence threshold, it is determined that the face does not exist in the to-be-detected image.

[0060] According to a third aspect of the embodiments of the present disclosure, an electronic device can include a processor and a memory for storing processor-executable instructions; and the processor is configured to execute the instructions to implement any of the optional image processing methods of the first aspect.

[0061] According to a fourth aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, and the computer-readable storage medium stores instructions, when the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device can execute any of the optional image processing methods of the first aspect.

[0062] According to a fifth aspect of the embodiments of the present disclosure, a computer program product is provided, which includes computer instructions, when the computer instructions are run on an electronic device, cause the electronic device to perform the image processing method according to any one of the optional implementation manners of the first aspect.

[0063] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure.

[0064] The technical solutions provided by the embodiments of the present disclosure at least have the following beneficial effects:

[0065] Based on any one of the above aspects, in the present disclosure, in the case that the first confidence output by the face detection model cannot accurately determine whether there is a face in the to-be-detected image, the electronic device can further determine whether there is a face in the to-be-detected image through the local detection model, which utilizes the combination of global information and local information of the face (for example, information of the whole face and local features), greatly improves the generalization ability of face detection, and improves the accuracy and recall rate of face detection. BRIEF DESCRIPTION OF DRAWINGS

[0066] The accompanying drawings, which are incorporated into and form part of the specification, illustrate embodiments consistent with the present disclosure and, together with the specification, serve to explain the principles of the present disclosure, and do not constitute an undue limitation on the present disclosure.

[0067] Figure 1 A flowchart of an image processing method provided by an embodiment of the present disclosure is shown;

[0068] Figure 2 A flowchart of another image processing method provided by an embodiment of the present disclosure is shown;

[0069] Figure 3 A flowchart of another image processing method provided by an embodiment of the present disclosure is shown;

[0070] Figure 4 A flowchart of another image processing method provided by an embodiment of the present disclosure is shown;

[0071] Figure 5 A flowchart of another image processing method provided by an embodiment of the present disclosure is shown;

[0072] Figure 6 A flowchart of another image processing method provided by an embodiment of the present disclosure is shown;

[0073] Figure 7 A flowchart of another image processing method provided by an embodiment of the present disclosure is shown;

[0074] Figure 8 Fig. 2 shows a flow diagram of another image processing method provided by an embodiment of the present disclosure;

[0075] Figure 9 Fig. 3 shows a flow diagram of another image processing method provided by an embodiment of the present disclosure;

[0076] Figure 10 Fig. 4 shows a flow diagram of another image processing method provided by an embodiment of the present disclosure;

[0077] Figure 11 Fig. 5 shows a structural diagram of an image processing device provided by an embodiment of the present disclosure;

[0078] Figure 12 Fig. 6 shows a structural diagram of a terminal provided by an embodiment of the present disclosure;

[0079] Figure 13 Fig. 7 shows a structural diagram of a server provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0080] In order to make the ordinary person skilled in the art better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be described clearly and completely below with reference to the drawings.

[0081] It should be noted that the terms "first", "second", and the like in the specification and claims of the present disclosure and the above-described drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or a chronological sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present disclosure described herein can be implemented in an order other than that illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Rather, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0082] It should also be understood that the term "comprising" indicates the presence of described features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, and / or components.

[0083] The data involved in the present disclosure can be data authorized by the user or sufficiently authorized by all parties.

[0084] In general technology, there is a certain degree of missed detection and false detection in face detection methods. Therefore, how to accurately identify the face in the image is a technical problem that needs to be solved at present.

[0085] Based on this, the embodiment of the present disclosure provides an image processing method. After obtaining a to-be-detected image, the to-be-detected image can be input into a face detection model to obtain a first confidence degree representing a probability of the presence of a face in the to-be-detected image and feature information of the to-be-detected image. Then, in the case that the first confidence degree is greater than a first confidence degree threshold and less than a second confidence degree threshold, the feature information of the to-be-detected image can be input into a local detection model to obtain local feature information and a second confidence degree, the second confidence degree representing a probability of the presence of a local feature of a face in an image output by the local detection model corresponding to the second confidence degree. Subsequently, whether a face is present in the to-be-detected image is determined according to a confidence degree set including the first confidence degree and the second confidence degree.

[0086] As can be seen from the above, in the case that the first confidence degree output by the face detection model cannot accurately determine whether a face is present in the to-be-detected image, the embodiment of the present disclosure can further determine whether a face is present in the to-be-detected image through the local detection model, and the combination of global information and local information of a face (for example, information of a whole face and local features of five facial features) is utilized, which greatly improves the generalization ability of face detection and improves the accuracy and recall rate of face detection.

[0087] The image processing method provided by the embodiment of the present disclosure is exemplarily described below.

[0088] The image processing method provided by the present disclosure can be applied to an electronic device.

[0089] In some embodiments, the electronic device can be a server, or a terminal, or other electronic devices used for image processing, and the present disclosure does not limit the same.

[0090] The server can be a single server, or a server cluster composed of multiple servers. In some embodiments, the server cluster can also be a distributed cluster. The present disclosure does not limit the specific implementation of the server.

[0091] The terminal can be a mobile phone, a tablet computer, a desktop computer, a laptop computer, a notebook computer, an ultra-mobile personal computer (UMPC), a netbook, a cellular phone, a personal digital assistant (PDA), an augmented reality (AR) \ virtual reality (VR) device, and other devices that can install and use a content community application (such as Kuaishou), and the present disclosure does not specially limit the specific form of the electronic device. It can interact with the user through one or more ways such as a keyboard, a touchpad, a touch screen, a remote control, voice interaction, or a handwriting device.

[0092] The image processing method provided by the embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings.

[0093] As shown in Figure 1 When the image processing method is applied to an electronic device, the image processing method can include:

[0094] S101, the electronic device obtains a to-be-detected image.

[0095] Optionally, the to-be-detected image can be sent by another electronic device, or can be obtained by the electronic device.

[0096] Optionally, the to-be-detected image can be a monitoring image taken in a large public area such as a shopping mall, a school, or a railway station, can be an image frame in a short video, or can be another type of image.

[0097] Optionally, the to-be-detected image obtained by the electronic device can be an original image, or can be an image obtained after pre-processing the original image.

[0098] The pre-processing can be size clipping processing, layer processing, definition processing, sharpening processing, etc.

[0099] S102, the electronic device inputs the to-be-detected image into a face detection model to obtain a first confidence and feature information of the to-be-detected image.

[0100] The first confidence represents the probability that a face exists in the to-be-detected image. The feature information of the to-be-detected image can include position information of key points in the to-be-detected image (such as position information of eyes, position information of a mouth, etc.). The position information can be represented using coordinates.

[0101] Specifically, after obtaining the to-be-detected image, the electronic device can first input the to-be-detected image into the face detection model to obtain the first confidence and the feature information of the to-be-detected image.

[0102] Optionally, the face detection model can be a single deep learning model, or can be an aggregation model obtained by fusing multiple different models.

[0103] The generalization ability of a single model is often relatively weak, and inevitably produces missed detection and false detection. The existing single model detection method cannot meet the requirements of high-precision face detection scenarios. The fusion model can improve the performance of machine learning by fusing multiple different models. The fusion model can often be fused from different angles such as model results, models themselves, and sample sets. Such methods can combine the advantages of multiple models to improve the prediction accuracy of the model and to some extent alleviate the shortcomings of the generalization ability of a single model. Therefore, the fusion model method is more and more widely used in scenarios where the precision and stability of a single model cannot meet the business requirements. The commonly used fusion model methods include weighted fusion method, stacking method (Stacking), bagging algorithm (Bagging) and boosting method (Boosting) and the like.

[0104] In an implementable manner, the face detection model in the present disclosure can be obtained by fusing a face detection sub-model, a face orientation model and a face key point detection model.

[0105] The face detection sub-model is used to determine the probability of the presence of a face in the to-be-detected image. The face orientation model is used to correct the face orientation. The face key point detection model is used to detect the face key points.

[0106] S103, in the case where the first confidence is greater than the first confidence threshold and less than the second confidence threshold, the electronic device inputs the feature information of the to-be-detected image into the local detection model to obtain local feature information and a second confidence.

[0107] The second confidence represents the probability of the presence of a face local feature in the feature information input into the local detection model.

[0108] Specifically, when the first confidence is greater than the first confidence threshold and less than the second confidence threshold, it indicates that the electronic device cannot accurately determine whether there is a face in the to-be-detected image through the face detection model. In this case, the electronic device can input the feature information of the to-be-detected image into the local detection model to obtain local feature information and a second confidence.

[0109] In an implementable manner, the local detection model includes at least one of a left eye detection model, a right eye detection model and a mouth detection model.

[0110] In the case where the local detection model includes the left eye detection model, the second confidence includes a left eye confidence output by the left eye detection model; the left eye confidence is used to represent the probability of the presence of a left eye feature in the left eye local feature information input into the left eye detection model.

[0111] And / or, in a case where the local detection model includes the right eye detection model, the second confidence includes a right eye confidence output by the right eye detection model; the right eye confidence is used to represent a probability that the right eye feature exists in the right eye local feature information input into the right eye detection model.

[0112] And / or, in a case where the local detection model includes the mouth detection model, the second confidence includes a mouth confidence output by the mouth detection model; the mouth confidence is used to represent a probability that the mouth feature exists in the mouth local feature information input into the mouth detection model.

[0113] In this way, after obtaining the to-be-detected image, if it is unable to accurately determine whether the face exists in the to-be-detected image by using the face detection model, at least one of the left eye detection model, the right eye detection model and the mouth detection model can be used to perform subsequent processing on the output result of the face detection model.

[0114] Since the left eye, the right eye and the mouth are used as inputs of the respective local detection models, compared with the general model fusion method, more obvious semantic information (local images of the left eye, the right eye and the mouth) is used, and therefore the disclosure has better generalization ability. By using the local information, the missed detection and the false detection samples generated by the face detection model can be effectively filtered, and therefore the accuracy and the recall rate of the face detection can be greatly improved.

[0115] In S104, the electronic device determines whether the face exists in the to-be-detected image according to the confidence set including the first confidence and the second confidence.

[0116] Specifically, after determining the first confidence and the second confidence, the electronic device can determine whether the face exists in the to-be-detected image according to the confidence set including the first confidence and the second confidence.

[0117] Optionally, the electronic device can determine a sum value of all the confidences in the confidence set, and determine whether the sum value is located in a preset range. When the sum value is located in the preset range, it is determined that the face exists in the to-be-detected image. Correspondingly, when the sum value is not located in the preset range, it is determined that the face does not exist in the to-be-detected image.

[0118] Optionally, the electronic device can also determine whether the face exists in the to-be-detected image according to the confidence set including the first confidence and the second confidence and other algorithms (for example, weighted summation, etc.).

[0119] The technical solutions provided by the above embodiments bring at least the following beneficial effects: As can be seen from S101-S104, this disclosure can further determine whether there is a face in the image to be detected by using a local detection model when the first confidence level output by the face detection model cannot accurately determine whether there is a face in the image to be detected. It utilizes the combination of global and local information of the face (such as information of the whole face and the local features), which greatly improves the generalization ability of face detection and improves the accuracy and recall rate of face detection.

[0120] In one embodiment, combined with Figure 1 ,like Figure 2 As shown, when the face detection model includes: a face detection sub-model for determining the probability of a face existing in the image to be detected, a face orientation model for correcting the face orientation, and a face keypoint detection model for detecting face keypoints, the feature information of the image to be detected also includes the position information of the face keypoints in the image to be detected. In this case, the method in S102 above, where the electronic device inputs the image to be detected into the face detection model to obtain the first confidence score and the feature information of the image to be detected, specifically includes:

[0121] S201. The electronic device inputs the image to be detected into the face detection sub-model to obtain face boundary information and the first confidence level.

[0122] For example, such as Figure 3 As shown, the input of the face detection sub-model is the image to be detected (which may contain a face or not), and the output is the face boundary information and the first confidence score.

[0123] Optionally, the face boundary information can be the coordinates of the face bounding box.

[0124] S202. The electronic device inputs the face boundary information into the face orientation model to obtain the corrected image to be detected.

[0125] Since the orientation of the face in the image to be detected is not upright, the electronic device can input the face boundary information into the face orientation model to obtain a corrected image to be detected.

[0126] S203. The electronic device inputs the corrected image to be detected into the facial landmark detection model to obtain the third confidence level and facial landmark location information.

[0127] The third confidence level represents the probability that a human face exists in the corrected image to be detected.

[0128] Optionally, facial key points can be the eyes or the mouth of the face.

[0129] For example, such asFigure 4 As shown, after obtaining the to-be-detected image 401, the electronic device can input the to-be-detected image 401 into the face detection sub-model 402 to obtain the first confidence indicating the probability that a face exists in the to-be-detected image and the feature information of the to-be-detected image. Then, the electronic device can input the face boundary information into the face orientation model 403 to obtain the corrected to-be-detected image. Then, the electronic device inputs the corrected to-be-detected image into the face key point model 404 to obtain the third confidence and the face key point position information. In this case, the electronic device can generate the target image 405 including the face according to the face key point position information.

[0130] The technical solutions provided by the above embodiments at least have the following beneficial effects: As can be seen from S201-S203, when the electronic device inputs the to-be-detected image into the face detection model to obtain the first confidence and the feature information of the to-be-detected image, the face detection sub-model, the face orientation model, and the face key point detection model can be used to accurately determine the corresponding key point position information and further output a more accurate face image.

[0131] In an embodiment, in combination with Figure 2 As shown in (a) of FIG. 4, when the local feature information further includes the local key point position information and the occlusion state of the local key point, the occlusion state is unoccluded or occluded. In S103, the method for inputting the feature information of the to-be-detected image into the local detection model to obtain the local feature information and the second confidence includes: Figure 5 S501, when the face key point position information includes the left eye position information, input the left eye position information into the left eye detection model to obtain the left eye confidence, the left eye key point position information in the left eye local feature information, and the left eye key point occlusion state.

[0132] As shown in (a) of FIG. 4, the input of the left eye detection model is the left eye position information in the face key point position information, and the output is the left eye confidence, the left eye key point position information in the left eye local feature information, and the left eye key point occlusion state.

[0133] Figure 6 S502, when the face key point position information includes the right eye position information, input the right eye position information into the right eye detection model to obtain the right eye confidence, the right eye key point position information in the right eye local feature information, and the right eye key point occlusion state.

[0134] As shown in (a) of FIG. 4, the input of the left eye detection model is the left eye position information in the face key point position information, and the output is the left eye confidence, the left eye key point position information in the left eye local feature information, and the left eye key point occlusion state.

[0135] As shown in (a) of FIG. 4, the input of the left eye detection model is the left eye position information in the face key point position information, and the output is the left eye confidence, the left eye key point position information in the left eye local feature information, and the left eye key point occlusion state. Figure 6 ​As shown in (b), the input of the right eye detection model is the right eye position information in the facial key point position information, and the output is the right eye confidence, the right eye key point position information in the right eye local feature information, and the right eye key point occlusion status.

[0136] S503. When the facial key point location information includes mouth location information, the mouth location information is input into the mouth detection model to obtain the mouth confidence, the mouth key point location information in the mouth local feature information, and the mouth key point occlusion state.

[0137] For example, such as Figure 6 As shown in (c), the input of the mouth detection model is the mouth position information in the facial key point position information, and the output is the mouth confidence score, the mouth key point position information in the local mouth feature information, and the mouth key point occlusion status.

[0138] Optionally, local keypoints can be keypoints in a local location.

[0139] For example, eye keypoints can be multiple keypoints around the eyes. Mouth keypoints can be multiple keypoints around the mouth.

[0140] It should be noted that the order in which the electronic device executes S501, S502 and S503 is not limited.

[0141] The technical solution provided by the above embodiments brings at least the following beneficial effects: As shown in S501-S503, each local detection model can also output corresponding key point location information and occlusion status. In this way, this disclosure can further output more accurate face images.

[0142] In one embodiment, combined with Figure 5 ,like Figure 7 As shown, the confidence set also includes a third confidence level; in S104 above, the method by which the electronic device determines whether a face exists in the image to be detected based on the confidence set including the first confidence level and the second confidence level specifically includes:

[0143] S701. The electronic device determines the target value based on the first confidence level, the second confidence level, the third confidence level, and the preset algorithm.

[0144] Optionally, the preset algorithm may be a weighted sum of the first confidence level, the second confidence level, and the third confidence level, or other algorithms, which are not limited in this disclosure.

[0145] S702. When the target value is greater than or equal to the preset threshold, the electronic device determines that there is a human face in the image to be detected.

[0146] S703, when the target value is less than the preset threshold, the electronic device determines that the face does not exist in the to-be-detected image.

[0147] The technical solutions provided by the above embodiments at least have the following beneficial effects: as can be seen from S701-S703, the electronic device can quickly and accurately determine whether the face exists in the to-be-detected image according to the first confidence, the second confidence, the third confidence, and the preset algorithm, thereby improving the efficiency of face detection.

[0148] In an embodiment, in combination with Figure 7 As Figure 8 indicated, the image processing method further includes:

[0149] S801, when the first confidence is greater than or equal to the second confidence threshold, the electronic device determines that the face exists in the to-be-detected image, and outputs the feature information of the to-be-detected image.

[0150] Specifically, when the first confidence is greater than or equal to the second confidence threshold, it indicates that the electronic device can accurately determine that the face exists in the to-be-detected image through the face detection model. In this case, the electronic device can determine that the face exists in the to-be-detected image, and output the feature information of the to-be-detected image.

[0151] S802, when the first confidence is less than or equal to the first confidence threshold, the electronic device determines that the face does not exist in the to-be-detected image.

[0152] Specifically, when the first confidence is less than or equal to the first confidence threshold, it indicates that the electronic device can accurately determine that the face does not exist in the to-be-detected image through the face detection model. In this case, the electronic device outputs the detection result that the face does not exist in the to-be-detected image.

[0153] For example, as Figure 9 indicated, after obtaining the to-be-detected image 901, the electronic device can input the to-be-detected image 901 into the face detection model 902 to obtain the first confidence.

[0154] When the first confidence is greater than or equal to the second confidence threshold, the electronic device determines that the face exists in the to-be-detected image 901, and outputs the feature information 903 of the to-be-detected image.

[0155] When the first confidence is greater than the first confidence threshold and less than the second confidence threshold, the electronic device inputs the to-be-detected image into the local detection model 904 to obtain the local feature information 905 and the second confidence. Subsequently, the electronic device can determine whether the face exists in the to-be-detected image according to the confidence set including the first confidence and the second confidence.

[0156] When the first confidence is greater than or equal to the second confidence threshold, the electronic device determines that the face does not exist in the to-be-detected image 901.

[0157] Through the confidence threshold classification method, the disclosure can determine that the to-be-detected image with high confidence contains a face, the to-be-detected image with low confidence does not contain a face, and the to-be-detected image with confidence between the two confidence thresholds is further screened by a subsequent local detection model. Through the double-threshold division strategy, the to-be-detected image with uncertain face detection sub-model (the first confidence is between the two confidence thresholds) is further screened, which can effectively reduce the probability of missed detection and missed detection, thereby improving the accuracy and recall of face detection.

[0158] The technical solutions provided by the above embodiments at least have the following beneficial effects: as can be known from 801-S802, when the first confidence output by the face detection model can accurately determine whether a face exists in the to-be-detected image, the electronic device can directly output the result, thereby improving the face detection efficiency.

[0159] In an embodiment, as shown in Figure 10 The image processing method includes:

[0160] S1001, the electronic device inputs a to-be-detected image into a face detection sub-model to obtain face boundary information and a first confidence.

[0161] Optionally, the size of the to-be-detected image can be 3x256x256.

[0162] When the first confidence is greater than or equal to the second confidence threshold, the electronic device performs S1002-S1003; when the first confidence is greater than the first confidence threshold and less than the second confidence threshold, the electronic device performs S1004-S1011; and when the first confidence is greater than or equal to the second confidence threshold, the electronic device performs S1012.

[0163] Optionally, after obtaining the face boundary information, the face region can be cropped according to the face boundary information. In order to facilitate subsequent model processing, the size of the image obtained after cropping can be adjusted to 3x256x256 to obtain an image containing only face boundary information, which is used as the input of the face orientation model.

[0164] S1002, the electronic device inputs the face boundary information into a face orientation model to obtain a corrected to-be-detected image.

[0165] S1003, the electronic device inputs the corrected to-be-detected image into a face key point detection model to obtain face key point position information and outputs a detection result.

[0166] The detection result includes face key point position information and a face existing in the to-be-detected image.

[0167] Optionally, the number of face key points can be 101.

[0168] In S1004, the electronic device inputs the face boundary information into the face orientation model to obtain a corrected to-be-detected image.

[0169] In S1005, the electronic device inputs the corrected to-be-detected image into the face key point detection model to obtain third confidence and face key point position information.

[0170] Optionally, the number of face key points can be 101.

[0171] Optionally, the electronic device can also crop a left eye image according to the left eye key point information, and adjust the size of the cropped image to 3x256x256 to obtain a local image containing only the left eye.

[0172] Correspondingly, the electronic device can also crop a right eye image according to the right eye key point information, and adjust the size of the cropped image to 3x256x256 to obtain a local image containing only the right eye.

[0173] Correspondingly, the electronic device can also crop a mouth image according to the mouth key point information, and adjust the size of the cropped image to 3x256x256 to obtain a local image containing only the mouth.

[0174] In S1006, when the face key point position information includes left eye position information, the left eye position information is input into a left eye detection model to obtain left eye confidence, left eye key point position information in left eye local feature information, and left eye key point occlusion state.

[0175] In S1007, when the face key point position information includes right eye position information, the right eye position information is input into a right eye detection model to obtain right eye confidence, right eye key point position information in right eye local feature information, and right eye key point occlusion state.

[0176] In S1008, when the face key point position information includes mouth position information, the mouth position information is input into a mouth detection model to obtain mouth confidence, mouth key point position information in mouth local feature information, and mouth key point occlusion state.

[0177] In S1009, the electronic device determines a target value according to the first confidence, the second confidence, the third confidence, and a preset algorithm.

[0178] S1010, when the target value is greater than or equal to a preset threshold, the electronic device determines that the face exists in the to-be-detected image.

[0179] S1011, when the target value is less than the preset threshold, the electronic device determines that the face does not exist in the to-be-detected image.

[0180] S1012, when the first confidence is less than or equal to the first confidence threshold, the electronic device determines that the face does not exist in the to-be-detected image.

[0181] It can be understood that, in actual implementation, the terminal / server described in the embodiments of the present disclosure can contain one or more hardware structures and / or software modules for implementing the foregoing corresponding image processing method, and these execution hardware structures and / or software modules can constitute an electronic device. Those skilled in the art should easily realize that, in combination with the algorithm steps of each example described in the embodiments disclosed herein, the present disclosure can be realized in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed in the form of hardware or computer software driven hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered beyond the scope of the present disclosure.

[0182] Based on such understanding, the embodiments of the present disclosure also correspondingly provide an image processing apparatus, which can be applied to an electronic device. Figure 11 The structure schematic diagram of the image processing apparatus provided by the embodiments of the present disclosure is shown. As shown in the figure, the image processing apparatus can include an acquisition unit 1101 and a processing unit 1102. Figure 11 The acquisition unit 1101 is configured to acquire a to-be-detected image.

[0183] The processing unit 1102 is configured to input the to-be-detected image into a face detection model to obtain a first confidence and feature information of the to-be-detected image; the first confidence represents a probability that a face exists in the to-be-detected image.

[0184] The processing unit 1102 is further configured to, in a case where the first confidence is greater than a first confidence threshold and less than a second confidence threshold, input the feature information of the to-be-detected image into a local detection model to obtain local feature information and a second confidence; the second confidence represents a probability that a local feature of a face exists in the feature information input into the local detection model.

[0185] The processing unit 1102 is further configured to determine whether the face exists in the to-be-detected image according to a confidence set including the first confidence and the second confidence.

[0186] The processing unit 1102 is further configured to determine whether the face exists in the to-be-detected image according to a confidence set including the first confidence and the second confidence.

[0187] Optionally, the local detection model comprises at least one of a left eye detection model, a right eye detection model, and a mouth detection model.

[0188] In a case where the local detection model comprises the left eye detection model, the second confidence comprises a left eye confidence output by the left eye detection model; the left eye confidence is used to represent a probability that the left eye feature exists in the left eye local feature information input into the left eye detection model.

[0189] In a case where the local detection model comprises the right eye detection model, the second confidence comprises a right eye confidence output by the right eye detection model; the right eye confidence is used to represent a probability that the right eye feature exists in the right eye local feature information input into the right eye detection model.

[0190] In a case where the local detection model comprises the mouth detection model, the second confidence comprises a mouth confidence output by the mouth detection model; the mouth confidence is used to represent a probability that the mouth feature exists in the mouth local feature information input into the mouth detection model.

[0191] Optionally, the face detection model comprises a face detection sub-model used to determine a probability that a face exists in the to-be-detected image, a face orientation model used to correct a face orientation, and a face key point detection model used to detect face key points; the feature information of the to-be-detected image further comprises face key point position information in the to-be-detected image.

[0192] The processing unit 1102 is specifically configured to:

[0193] input the to-be-detected image into the face detection sub-model to obtain face boundary information and a first confidence;

[0194] input the face boundary information into the face orientation model to obtain a corrected to-be-detected image;

[0195] input the corrected to-be-detected image into the face key point detection model to obtain third confidence and face key point position information; the third confidence represents a probability that the face exists in the corrected to-be-detected image.

[0196] Optionally, the local feature information further comprises local key point position information and a local key point occlusion state; the occlusion state is unoccluded or occluded.

[0197] The processing unit 1102 is specifically configured to:

[0198] when the face key point position information comprises left eye position information, input the left eye position information into the left eye detection model to obtain left eye confidence, left eye key point position information in the left eye local feature information, and a left eye key point occlusion state;

[0199] And / or, when the face key point position information comprises right eye position information, inputting the right eye position information into a right eye detection model to obtain right eye confidence, right eye key point position information in right eye local feature information, and a right eye key point occlusion state.

[0200] And / or, when the face key point position information comprises mouth position information, inputting the mouth position information into a mouth detection model to obtain mouth confidence, mouth key point position information in mouth local feature information, and a mouth key point occlusion state.

[0201] Optionally, the confidence set further comprises a third confidence.

[0202] The processing unit 1102 is specifically configured to:

[0203] determine a target value according to the first confidence, the second confidence, the third confidence, and a preset algorithm;

[0204] When the target value is greater than or equal to a preset threshold, it is determined that the face exists in the to-be-detected image.

[0205] Or, when the target value is less than the preset threshold, it is determined that the face does not exist in the to-be-detected image.

[0206] Optionally, the processing unit 1102 is further configured to:

[0207] When the first confidence is greater than or equal to a second confidence threshold, it is determined that the face exists in the to-be-detected image, and the feature information of the to-be-detected image is output.

[0208] Or, when the first confidence is less than or equal to a first confidence threshold, it is determined that the face does not exist in the to-be-detected image.

[0209] As described above, the embodiments of the present disclosure can divide the functional modules of the electronic device according to the above-mentioned method examples. Among them, the integrated modules can be realized in the form of hardware or in the form of software functional modules. In addition, it needs to be explained that the division of modules in the embodiments of the present disclosure is illustrative, and is only a logical function division. When actually implemented, there can be another division method. For example, each functional module can be divided according to each function, or two or more functions can be integrated into one processing module.

[0210] As for the image processing apparatus in the above-mentioned embodiments, the specific manner in which each module performs operations and the beneficial effects possessed have been described in detail in the foregoing method embodiments, which will not be described here again.

[0211] The embodiments of the present disclosure also provide a terminal, which can be a mobile phone, a computer, or other user terminals. Figure 12A structure schematic diagram of a terminal provided by an embodiment of the present disclosure is shown. The terminal can be an image processing apparatus and can include at least one processor 61, a communication bus 62, a memory 63, and at least one communication interface 64.

[0212] The processor 61 can be a central processing unit (CPU), a micro-processing unit, an ASIC, or one or more integrated circuits for controlling program execution of the present disclosure. As an example, in combination with the present disclosure, the processor 61 can be implemented by a processing unit 1102 in an electronic device. Figure 11 The processing unit 1102 in the electronic device implements the same functions as the processor 61 in the image processing apparatus. Figure 12

[0213] The communication bus 62 can include a path for transmitting information between the above-mentioned components.

[0214] The communication interface 64 uses any transceiver-like device for communicating with other devices or communication networks, such as a server, an Ethernet, a radio access network (RAN), a wireless local area networks (WLAN), etc. As an example, the communication interface 64 can include a modem, a network interface card (NIC), or other well-known interface devices for connecting to other devices or networks.

[0215] The memory 63 can be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disk storage, a magnetic disk storage or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to this. The memory can exist independently and be connected to the processing unit through a bus. The memory can also be integrated with the processing unit.

[0216] The memory 63 is used to store application program codes for executing the present disclosure, and the processor 61 is used to control the execution. The processor 61 is used to execute the application program codes stored in the memory 63, thereby realizing the functions in the present disclosure.

[0217] ​In a specific implementation, as one example, processor 61 may include one or more CPUs, for example... Figure 12 CPU0 and CPU1 in the CPU.

[0218] In a specific implementation, as one example, the terminal may include multiple processors, for example... Figure 12 Processors 61 and 65 are included. Each of these processors may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. A processor here may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0219] In a specific implementation, as one embodiment, the terminal may further include an input device 66 and an output device 67. The input device 66 and output device 67 communicate and can accept user input in various ways. For example, the input device 66 may be a mouse, keyboard, touchscreen device, or sensing device. The output device 67 communicates with the processor 61 and can display information in various ways. For example, the output device 61 may be a liquid crystal display (LCD), a light emitting diode (LED) display device, etc.

[0220] Those skilled in the art will understand that Figure 12 The structure shown does not constitute a limitation on the terminal and may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0221] This disclosure also provides a server. Figure 13 A schematic diagram of the server structure provided in an embodiment of this disclosure is shown. This server can be an image processing device. The server can vary significantly due to differences in configuration or performance, and may include one or more processors 71 and one or more memories 72. The memory 72 stores at least one instruction, which is loaded and executed by the processor 71 to implement the image processing methods provided in the various method embodiments described above. Of course, the server may also have wired or wireless network interfaces, a keyboard, and input / output interfaces for input and output. The server may also include other components for implementing device functions, which will not be elaborated upon here.

[0222] The present disclosure also provides a computer readable storage medium including instructions, the computer readable storage medium having instructions stored thereon, which when executed by a processor of a computer device, enable the computer to perform the image processing method provided by the above-mentioned embodiments. For example, the computer readable storage medium can be a memory 63 including instructions, which can be executed by the processor 61 of the terminal to complete the above-mentioned method. For another example, the computer readable storage medium can be a memory 72 including instructions, which can be executed by the processor 71 of the server to complete the above-mentioned method. Alternatively, the computer readable storage medium can be a non-transitory computer readable storage medium, for example, the non-transitory computer readable storage medium can be a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0223] The present disclosure also provides a computer program product, which includes computer instructions, when the computer instructions run on an electronic device, enable the electronic device to perform the image processing method shown in any of the accompanying drawings. Figures 1-10 The present disclosure also provides a computer program product, which includes computer instructions, when the computer instructions run on an electronic device, enable the electronic device to perform the image processing method shown in any of the accompanying drawings.

[0224] Other embodiments of the present disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the features disclosed herein. The present disclosure is intended to cover any variations, uses or adaptations of the present disclosure following the general principles thereof and including such departures from the present disclosure that come within known

[0225] It should be understood that the present disclosure is not limited to the precise structures as has been described above and shown in the accompanying drawings, and that various modifications and changes can be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. An image processing method, characterized by, The method comprises: obtaining a to-be-detected image; inputting the to-be-detected image into a face detection model to obtain a first confidence and feature information of the to-be-detected image; the first confidence represents a probability that a face exists in the to-be-detected image; in a case where the first confidence is greater than a first confidence threshold and less than a second confidence threshold, inputting the feature information of the to-be-detected image into a local detection model to obtain local feature information and a second confidence; the second confidence represents a probability that a local feature of a face exists in the feature information input into the local detection model; the local detection model comprises at least one of a left eye detection model, a right eye detection model, and a mouth detection model; determining whether a face exists in the to-be-detected image according to a confidence set comprising the first confidence and the second confidence; wherein the local feature information further comprises local key point position information and a blocking state of a local key point; the blocking state is unblocked or blocked; the inputting the feature information of the to-be-detected image into the local detection model to obtain the local feature information and the second confidence comprises: when the face key point position information comprises left eye position information, inputting the left eye position information into the left eye detection model to obtain a left eye confidence, left eye key point position information in left eye local feature information, and a left eye key point blocking state; and / or, when the face key point position information comprises right eye position information, inputting the right eye position information into the right eye detection model to obtain a right eye confidence, right eye key point position information in right eye local feature information, and a right eye key point blocking state; and / or, when the face key point position information comprises mouth position information, inputting the mouth position information into the mouth detection model to obtain a mouth confidence, mouth key point position information in mouth local feature information, and a mouth key point blocking state.

2. The image processing method of claim 1, wherein, in a case where the local detection model comprises the left eye detection model, the second confidence comprises a left eye confidence output by the left eye detection model; the left eye confidence is used to represent a probability that a left eye feature exists in the left eye local feature information input into the left eye detection model; and / or, in a case where the local detection model comprises the right eye detection model, the second confidence comprises a right eye confidence output by the right eye detection model; the right eye confidence is used to represent a probability that a right eye feature exists in the right eye local feature information input into the right eye detection model; and / or, in a case where the local detection model comprises the mouth detection model, the second confidence comprises a mouth confidence output by the mouth detection model; the mouth confidence is used to represent a probability that a mouth feature exists in the mouth local feature information input into the mouth detection model.

3. The image processing method of claim 2, wherein, The face detection model comprises a face detection sub-model for determining a probability of presence of a face in the image to be detected, a face orientation model for correcting a face orientation, and a face key point detection model for detecting face key points; and the feature information of the image to be detected further comprises face key point position information in the image to be detected. The inputting of the image to be detected into the face detection model to obtain the first confidence and the feature information of the image to be detected comprises: inputting the image to be detected into the face detection sub-model to obtain face boundary information and the first confidence; inputting the face boundary information into the face orientation model to obtain a corrected image to be detected; inputting the corrected image to be detected into the face key point detection model to obtain a third confidence and the face key point position information; and the third confidence represents a probability of presence of a face in the corrected image to be detected.

4. The image processing method of claim 3, wherein, The confidence set further comprises the third confidence. The determination of whether a face is present in the image to be detected according to the confidence set comprising the first confidence and the second confidence comprises: determining a target value according to the first confidence, the second confidence, the third confidence, and a preset algorithm; when the target value is greater than or equal to a preset threshold, determining that a face is present in the image to be detected; or, when the target value is less than the preset threshold, determining that a face is not present in the image to be detected.

5. The image processing method according to any one of claims 1-4, characterized in that, Further comprising: when the first confidence is greater than or equal to the second confidence threshold, determining that a face is present in the image to be detected, and outputting the feature information of the image to be detected; or, when the first confidence is less than or equal to the first confidence threshold, determining that a face is not present in the image to be detected.

6. An image processing apparatus characterized by comprising: comprising: an acquisition unit and a processing unit; the acquisition unit is configured to acquire an image to be detected; the processing unit is configured to input the image to be detected into a face detection model to obtain a first confidence and feature information of the image to be detected; the first confidence represents a probability of presence of a face in the image to be detected; the processing unit is further configured to, in a case where the first confidence is greater than a first confidence threshold and less than a second confidence threshold, input the feature information of the image to be detected into a local detection model to obtain local feature information and a second confidence; the second confidence represents a probability of presence of a local face feature in the feature information input into the local detection model; the local detection model comprises at least one of a left eye detection model, a right eye detection model, and a mouth detection model; the processing unit is further configured to determine whether a face is present in the image to be detected according to a confidence set comprising the first confidence and the second confidence; wherein the local feature information further comprises local key point position information and a local key point occlusion state; the occlusion state is unoccluded or occluded; and the processing unit is specifically configured to: when the face key point position information comprises left eye position information, inputting the left eye position information into the left eye detection model to obtain left eye confidence, left eye key point position information in left eye local feature information, and left eye key point occlusion state; and / or, when the face key point position information comprises right eye position information, inputting the right eye position information into the right eye detection model to obtain right eye confidence, right eye key point position information in right eye local feature information, and right eye key point occlusion state; and / or, when the face key point position information comprises mouth position information, inputting the mouth position information into the mouth detection model to obtain mouth confidence, mouth key point position information in mouth local feature information, and mouth key point occlusion state.

7. The image processing apparatus of claim 6, wherein, when the local detection model comprises the left eye detection model, the second confidence comprises left eye confidence output by the left eye detection model, the left eye confidence being used to indicate a probability that left eye feature exists in left eye local feature information input into the left eye detection model; and / or, when the local detection model comprises the right eye detection model, the second confidence comprises right eye confidence output by the right eye detection model, the right eye confidence being used to indicate a probability that right eye feature exists in right eye local feature information input into the right eye detection model; and / or, when the local detection model comprises the mouth detection model, the second confidence comprises mouth confidence output by the mouth detection model, the mouth confidence being used to indicate a probability that mouth feature exists in mouth local feature information input into the mouth detection model.

8. The image processing apparatus according to claim 7, characterized by the face detection model comprises a face detection sub-model used to determine a probability that a face exists in the image to be detected, a face orientation model used to correct face orientation, and a face key point detection model used to detect face key points; and the feature information of the image to be detected further comprises face key point position information of the image to be detected; the processing unit is specifically configured to: input the image to be detected into the face detection sub-model to obtain face boundary information and the first confidence; input the face boundary information into the face orientation model to obtain a corrected image to be detected; input the corrected image to be detected into the face key point detection model to obtain third confidence and the face key point position information; the third confidence indicates a probability that a face exists in the corrected image to be detected.

9. The image processing apparatus according to claim 8, characterized by, the confidence set further comprises the third confidence; the processing unit is specifically configured to: determine a target value according to the first confidence, the second confidence, the third confidence, and a preset algorithm; when the target value is greater than or equal to a preset threshold, determine that a face exists in the image to be detected; or, when the target value is less than the preset threshold, determine that a face does not exist in the image to be detected.

10. The image processing apparatus according to any one of claims 6-9, characterized by, the processing unit is further configured to: When the first confidence is greater than or equal to the second confidence threshold, it is determined that a face exists in the image to be detected, and feature information of the image to be detected is outputted. Or, when the first confidence is less than or equal to the first confidence threshold, it is determined that a face does not exist in the image to be detected.

11. An electronic device, comprising: The electronic device comprises: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement the image processing method of any one of claims 1-5.

12. A computer-readable storage medium having stored thereon instructions, the instructions comprising, When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device is enabled to perform the image processing method of any one of claims 1-5.

13. A computer program product comprising instructions, characterized in that, When the instructions run on the electronic device, the electronic device performs the image processing method of any one of claims 1-5.

Images