Pet identity verification method, device and equipment and storage medium

Abstract

The embodiment of the invention provides a pet identity verification method and device, computer equipment and a computer readable storage medium, and relates to the field of finance and artificial intelligence. Obtaining audio and video information of a to-be-detected pet, wherein the audio and video information comprises a pet video and a pet audio corresponding to the to-be-detected pet; performing feature extraction and recognition processing on the pet audio to obtain pet voiceprint features; performing feature extraction and recognition processing on the pet video to obtain pet appearance features; obtaining identity feature information of the target pet, wherein the identity feature information comprises a target voiceprint feature and a target appearance feature; determining the similarity between the pet voiceprint feature and the target voiceprint feature as a first similarity, and determining the similarity between the pet appearance feature and the target appearance feature as a second similarity; and generating an identity verification result of the to-be-detected pet and the target pet according to the first similarity and the second similarity. The invention aims to solve the problem that identity information of an insured pet and a claim settlement pet cannot be accurately identified.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular to a pet identity verification method, a pet identity verification device, a computer device, and a computer-readable storage medium. Background Art

[0002] As the number of pet owners continues to grow and the concept of pet ownership continues to improve, the pet economy will continue to flourish. The premium scale of pet medical insurance has been rising year by year, and the accident rate has also been increasing. Therefore, it is extremely important for insurance companies to identify the identity of pets when making claims, and there is a certain deviation if the claims adjuster only relies on the naked eye to judge the identity of the pet.

[0003] At present, the strategy adopted by insurance companies is mostly to only identify photos of pets' nose prints or frontal faces. That is, when users purchase pet insurance, they need to upload photos of the insured pet's nose prints or frontal faces, and when making a claim, they also need to upload photos of the pet's nose prints or frontal faces. The insurance company verifies whether the photos before and after are of the same insured pet. However, due to interference factors such as user equipment, the environment in which they are located, and the pet's own liveliness, it is difficult to accurately and clearly take photos of the pet's nose prints and frontal faces, which makes it impossible for insurance companies to accurately identify the identity information of the insured pet and the claiming pet. Summary of the invention

[0004] The present application provides a pet identity verification method, a pet identity verification device, a computer device and a computer-readable storage medium, aiming to solve the problem of being unable to accurately identify the identity information of the insured pet and the claim pet.

[0005] To achieve the above purpose, the present application provides a pet identity verification method, the method comprising:

[0006] Acquire audio and video information of the pet to be detected, wherein the audio and video information includes a pet video and a pet audio corresponding to the pet to be detected;

[0007] Extracting and identifying the pet audio to obtain the pet voiceprint feature;

[0008] Performing feature extraction and recognition processing on the pet video to obtain the pet's appearance features;

[0009] Acquire identity feature information of a target pet, wherein the identity feature information includes a target voiceprint feature and a target appearance feature;

[0010] Determine the similarity between the pet's voiceprint feature and the target voiceprint feature as a first similarity, and determine the similarity between the pet's appearance feature and the target appearance feature as a second similarity;

[0011] An identity verification result of the pet to be detected and the target pet is generated according to the first similarity and the second similarity.

[0012] To achieve the above objectives, the present application provides a pet identity verification device, the pet identity verification device comprising:

[0013] An audio and video acquisition module, used to acquire audio and video information of the pet to be detected, wherein the audio and video information includes a pet video and a pet audio corresponding to the pet to be detected;

[0014] A voiceprint feature extraction module, used to extract and identify the pet audio to obtain the pet voiceprint feature;

[0015] An appearance feature extraction module is used to extract features and perform recognition processing on the pet video to obtain the pet's appearance features;

[0016] An identity feature acquisition module, used to acquire identity feature information of a target pet, wherein the identity feature information includes a target voiceprint feature and a target appearance feature;

[0017] A feature comparison module, configured to determine the similarity between the pet's voiceprint feature and the target voiceprint feature as a first similarity, and to determine the similarity between the pet's appearance feature and the target appearance feature as a second similarity;

[0018] An identity recognition module is used to generate an identity verification result of the pet to be detected and the target pet according to the first similarity and the second similarity.

[0019] To achieve the above objectives, the present application provides a computer device, which includes a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program and implement the pet identity verification method described above when executing the computer program.

[0020] To achieve the above objectives, the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the processor implements the pet identity verification method as described above.

[0021] The pet identity verification method, pet identity verification device, computer equipment and computer-readable storage medium disclosed in the embodiments of the present application obtain audio and video information of the pet to be detected, the audio and video information includes the pet video and pet audio corresponding to the pet to be detected; extract and identify the pet audio to obtain the pet voiceprint feature; extract and identify the pet video to obtain the pet appearance feature; obtain the identity feature information of the target pet, the identity feature information includes the target voiceprint feature and the target appearance feature; determine the similarity between the pet voiceprint feature and the target voiceprint feature as the first similarity, and determine the similarity between the pet appearance feature and the target appearance feature as the second similarity; generate the identity verification result of the pet to be detected and the target pet according to the first similarity and the second similarity. In this way, the identity verification result of the pet to be detected and the target pet can be generated quickly and accurately, so that the falsification of pet images can be avoided, the risk of pet insurance can be strictly controlled, and the risk of insurance fraud can be greatly reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0023] Figure 1 This is a scenario diagram of a pet identity verification method provided by an embodiment of the present application;

[0024] Figure 2 It is a flowchart of a pet identity verification method provided in an embodiment of the present application;

[0025] Figure 3 is a schematic block diagram of a pet identity verification device provided in one embodiment of the present application;

[0026] Figure 4 It is a schematic block diagram of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0028] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor do they have to be executed in the order described. For example, some operations / steps can also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions. In addition, although the functional modules are divided in the device schematic, in some cases, the module division can be different from that in the device schematic.

[0029] The term "and / or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0030] At present, the strategy adopted by insurance companies is mostly to only identify photos of pets' nose prints or frontal faces. That is, when users purchase pet insurance, they need to upload photos of the insured pets' nose prints or frontal faces, and when making claims, they also need to upload photos of the pets' nose prints or frontal faces. The insurance company verifies whether the photos before and after belong to the same insured pet. However, due to interference factors such as user equipment, the environment in which they are located, and the pets' own liveliness and activity, it is difficult to accurately and clearly take photos of the pets' nose prints and frontal faces, which will have a great impact on the success rate of model recognition. Moreover, single image data is extremely easy to falsify, which poses a considerable challenge to the risk control of pet insurance. Therefore, in the context of the ever-expanding pet market, it is particularly important for insurance companies to develop a multimodal pet identity verification method to intercept claims risks and maintain the payout ratio at a healthy level.

[0031] The present application proposes a pet identity verification method, a pet identity verification device, a computer device and a computer-readable storage medium, which can quickly and accurately generate identity verification results of the pet to be detected and the target pet, thereby avoiding the falsification of pet images, strictly controlling the risks of pet insurance, and greatly reducing the risk of insurance fraud.

[0032] The method can be applied to a server, and of course, can also be applied to a terminal device to generate a corresponding image classification model, wherein the terminal device may include a fixed terminal such as a mobile phone, a tablet computer, a personal digital assistant (PDA), etc. The server can be, for example, a separate server or a server cluster. However, for ease of understanding, the following embodiments will be described in detail with a pet identity verification method applied to a server.

[0033] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0034] like Figure 1As shown, the pet identity verification method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. The application environment includes a terminal device 110 and a server 120, wherein the terminal device 110 can communicate with the server 120 through a network. Specifically, the server 120 obtains audio and video information of the pet to be detected, and the audio and video information includes the pet video and pet audio corresponding to the pet to be detected; performs feature extraction and recognition processing on the pet audio to obtain the pet voiceprint feature; performs feature extraction and recognition processing on the pet video to obtain the pet appearance feature; obtains the identity feature information of the target pet, and the identity feature information includes the target voiceprint feature and the target appearance feature; determines that the similarity between the pet voiceprint feature and the target voiceprint feature is the first similarity, and determines that the similarity between the pet appearance feature and the target appearance feature is the second similarity; generates the identity verification result of the pet to be detected and the target pet according to the first similarity and the second similarity, and sends the identity verification result to the terminal device 110. Among them, the server 120 can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms. The terminal device 110 can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, etc., but is not limited to this. The terminal and the server can be directly or indirectly connected via wired or wireless communication, and this application does not limit this.

[0035] See also Figure 2 , Figure 2 This is a schematic flow chart of a pet identity verification method provided by an embodiment of the present application. The pet identity verification method can quickly and accurately generate the identity verification results of the pet to be detected and the target pet, thereby avoiding pet image fraud, strictly controlling the risk of pet insurance, and greatly reducing the risk of insurance fraud.

[0036] like Figure 2 As shown, the pet identity verification method includes steps S101 to S106.

[0037] S101. Acquire audio and video information of a pet to be detected, where the audio and video information includes a pet video and pet audio corresponding to the pet to be detected.

[0038] The pet to be detected may be a pet that needs to be verified, such as a pet that needs to be compensated. The audio and video information may include pet videos and pet audio uploaded by users. Pet videos are videos that can show the whole body of the pet, and pet audio is audio that can show the pet's call.

[0039] For example, when a user needs to make a claim for a pet, the audio and video information of the pet to be claimed uploaded by the user is obtained. Before processing the audio and video information, it is possible to first detect whether the audio and video information includes pet video and pet audio. If it is determined that the audio and video information includes pet video and pet audio, the pet video and pet audio are subsequently processed.

[0040] S102: extract and identify pet audio features to obtain pet voiceprint features.

[0041] The pet voiceprint feature may include a voiceprint feature corresponding to the call of the pet to be detected.

[0042] In some embodiments, the pet audio is cropped and denoised to obtain optimized pet audio; based on the pet call feature extraction model, the optimized pet audio is feature extracted to obtain pet voiceprint features. In this way, pet voiceprint features can be quickly and accurately extracted from the pet audio, so as to subsequently identify the pet voiceprint features.

[0043] The optimized pet audio is the pet audio obtained after cropping and denoising. The pet call feature extraction model is obtained by training the ECAPA-TDNN model using the Additive Angular Margin Loss function as the loss function, and can extract pet voiceprint features from pet audio.

[0044] Specifically, data cropping is performed on the pet audio to crop out the pet audio with a high correlation with the pet's call, and then the pet audio is denoised using wavelet denoising to obtain the optimized pet audio; based on the pet's call feature extraction model, feature extraction is performed on the optimized pet audio to obtain the pet voiceprint feature.

[0045] Exemplarily, the call data and corresponding voiceprint data of multiple types of pets can be obtained, and the above data can be divided into a training set and a standard set. The training set and the standard set are input into the ECAPA-TDNN model for training. According to the data characteristics of the training set and the standard set, the additive angular margin loss function (Additive AngularMargin Loss) is used as the loss function to optimize the ECAPA-TDNN model, thereby training a pet call feature extraction model, which can quickly and accurately extract pet voiceprint features from pet audio.

[0046] S103: extracting features and performing recognition processing on the pet video to obtain the pet's appearance features.

[0047] The pet appearance features may include appearance features corresponding to the appearance of the pet to be detected.

[0048] In some embodiments, image detection processing is performed on the pet video to obtain a pet multi-dimensional image; target detection is performed on the pet multi-dimensional image to obtain pet object information corresponding to the pet multi-dimensional image; based on the pet appearance feature extraction model, feature extraction is performed on the pet object information corresponding to the pet multi-dimensional image to obtain pet appearance features. In this way, pet appearance features can be quickly and accurately extracted from the pet video, so as to subsequently identify the pet appearance features.

[0049] The pet multi-dimensional image may include a pet front face image, a pet side image and a pet top view image, which are obtained by identifying the pet video. The pet object information is the pet object intercepted from the pet multi-dimensional image, from which the pet outline and pet color and other information can be determined. The pet appearance feature extraction model is obtained by training the YOLOv8 model using the Additive Angular Margin Loss function as the loss function, and can extract the pet appearance features from the pet object information.

[0050] Specifically, the pet video can be processed through an image detection algorithm to select three high-definition dynamic dimensional images of the pet's front face, pet's side image and pet's overhead image as the pet's multi-dimensional image; target detection is performed on the pet's multi-dimensional image to obtain the pet object information corresponding to the pet's multi-dimensional image, such as the pet's outline and pet's color; based on the pet's appearance feature extraction model, feature extraction is performed on the pet object information corresponding to the pet's multi-dimensional image to obtain the pet's appearance features.

[0051] Exemplarily, multi-dimensional images of multiple types of pets and corresponding appearance feature data can be obtained, and the above data can be divided into a training set and a standard set, and the training set and the standard set are input into the YOLOv8 model for training. According to the data characteristics of the training set and the standard set, the additive angular margin loss function (Additive AngularMargin Loss) is used as the loss function to optimize the YOLOv8 model, thereby training a pet appearance feature extraction model, which can quickly and accurately extract pet appearance features from pet videos.

[0052] For example, the loss of the pet appearance feature extraction model for each dynamic dimension (frontal, sideways, and top view) is recorded as l face ,l side and l body Here, the additive angle margin loss function is used as the definition of the three loss functions. Then the final loss function definition of the training can be total=0.5l face +αl side +βl body (α+β=0.5). Regarding the external features of pets, the facial features of the front face are the most important basis for judging their appearance, especially their eye features, so l is assigned face At least half of the loss weight, the weight ratios α and β of the other two dynamic dimensions can be obtained through model training, and the weight ratio corresponding to each dynamic dimension (front face, side view, top view) can also be determined according to the image information.

[0053] It should be noted that the weight ratio can be determined according to image information or preset, and is not specifically limited here.

[0054] In some embodiments, feature extraction is performed on the pet object information corresponding to the pet front face image to obtain the pet front face feature; feature extraction is performed on the pet object information corresponding to the pet side body image to obtain the pet side body feature; feature extraction is performed on the pet object information corresponding to the pet top view image to obtain the pet top view feature; and the pet appearance feature is generated based on the pet front face feature, the pet side body feature and the pet top view feature. In this way, feature information corresponding to the pet front face image, the pet side body feature and the pet top view image can be extracted respectively, so as to accurately generate the pet appearance feature.

[0055] Among them, the pet's frontal facial features may include nose pattern features, facial contour features, eye features, etc. The pet's side profile features may include hair features, body shape features, etc. The pet's top view features may include hair features, body shape features, etc.

[0056] Exemplarily, feature extraction is performed on the pet object information corresponding to the pet's frontal face image to obtain pet frontal face features such as nose print features, facial contour features, eye features, etc.; feature extraction is performed on the pet object information corresponding to the pet's side view image to obtain pet side view features such as hair features, body shape features, etc.; feature extraction is performed on the pet object information corresponding to the pet's overhead view image to obtain pet overhead view features such as hair features, body shape features, etc.; finally, the pet's frontal face features, pet's side view features and pet's overhead view features are combined to generate the pet's appearance features.

[0057] In some embodiments, image information corresponding to the frontal face image, the side image and the overhead view image of the pet is obtained, and the image information includes image clarity and target object ratio; the weight ratio corresponding to the frontal face image, the side image and the overhead view image of the pet is determined according to the image clarity and the target object ratio; the pet appearance features are generated according to the frontal face features, the side features and the overhead view features of the pet and their corresponding weight ratios. Thus, the weight ratio corresponding to the frontal face image, the side image and the overhead view image of the pet can be accurately determined through the image information, so as to accurately generate the pet appearance features.

[0058] The image information may include information such as image clarity and target object ratio of the pet's front face image, pet's side image, and pet's overhead image. Image clarity is used to indicate the clarity of the pet in the image, which may affect the subsequent recognition results. The target object ratio is used to indicate the area ratio of the pet object in the image, which may affect the subsequent recognition results.

[0059] Specifically, the corresponding weight ratios of the pet's front face image, pet's side image and pet's overhead view image can be determined according to the image clarity and the target object ratio. Generally speaking, the higher the image clarity and the higher the target object ratio, the higher the weight ratio of the corresponding image.

[0060] Exemplarily, weight ratios of 60%, 30% and 10% can be assigned to the pet's front face image, pet's side view image and pet's overhead view image respectively according to their image clarity and target object ratio.

[0061] For example, assuming that the image clarity of the pet's front face image is 300DPI, the image clarity of the pet's side image is 100DPI, and the image clarity of the pet's overhead view image is 100DPI, then the first weight ratios corresponding to the pet's front face image, the pet's side image, and the pet's overhead view image can be determined to be 60%, 20%, and 20%; if the target object ratio of the pet's front face image is 100%, the target object ratio of the pet's side image is 60%, and the target object ratio of the pet's overhead view image is 40%, then the second weight ratios corresponding to the pet's front face image, the pet's side image, and the pet's overhead view image can be determined to be 50%, 30%, and 20%, so that the weighted sum of the first weight ratio and the second weight ratio can obtain the weight ratios of the pet's front face image, the pet's side image, and the pet's overhead view image of 55%, 25%, and 20%.

[0062] It should be noted that when the image clarity of the pet's front face image, pet's side view image and pet's overhead view image all exceed the clarity threshold, the corresponding weight ratios of the pet's front face image, pet's side view image and pet's overhead view image can be determined only by the target object ratio; when the target object ratios of the pet's front face image, pet's side view image and pet's overhead view image all exceed the object ratio threshold, the corresponding weight ratios of the pet's front face image, pet's side view image and pet's overhead view image can be determined only by the image clarity.

[0063] Preferably, with respect to pet appearance features, the pet's frontal face image is the most important basis for judging the pet's appearance features, especially their eye features, so the pet's frontal face image is generally given at least half of the weight ratio.

[0064] S104, obtaining identity feature information of the target pet, where the identity feature information includes target voiceprint features and target appearance features.

[0065] The target pet is a pet that needs to be identified, such as a pet that needs to be insured. The identity feature information includes a target voiceprint feature and a target appearance feature. The target voiceprint feature may be a voiceprint feature corresponding to the target pet, and the target appearance feature may be an appearance feature corresponding to the target pet.

[0066] For example, when a user needs to insure a pet, the audio and video information of the pet at the time of insurance is obtained, and then the audio of the pet at the time of insurance is feature extracted and recognized to obtain the target voiceprint feature, and then the video of the pet at the time of insurance is feature extracted and recognized to obtain the target appearance feature, and the target voiceprint feature and target appearance feature are saved for use in subsequent verification.

[0067] S105: Determine the similarity between the pet's voiceprint feature and the target's voiceprint feature as a first similarity, and determine the similarity between the pet's appearance feature and the target's appearance feature as a second similarity.

[0068] The first similarity may be the cosine similarity between the pet voiceprint feature and the target voiceprint feature, and the second similarity may be the cosine similarity between the pet appearance feature and the target appearance feature.

[0069] Specifically, the cosine similarity between the pet voiceprint feature and the target voiceprint feature can be calculated to accurately determine the first similarity; the cosine similarity between the pet appearance feature and the target appearance feature can be calculated to accurately determine the second similarity.

[0070] S106: Generate an identity verification result of the pet to be detected and the target pet according to the first similarity and the second similarity.

[0071] The identity verification results of the pet to be detected and the target pet may include two results: the identities of the pet to be detected and the target pet are the same; and the identities of the pet to be detected and the target pet are different.

[0072] In some embodiments, the pet claim risk level is obtained; if the pet claim risk level is the first level, when the first similarity and the second similarity both exceed the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet; if the pet claim risk level is the second level, when the first similarity or the second similarity exceeds the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet. In this way, the identity information of the pet to be detected and the target pet can be accurately determined, thereby avoiding pet image fraud, strictly controlling the risk of pet insurance, and greatly reducing the risk of insurance fraud.

[0073] The pet claim risk level may include a first level and a second level, the first level may be used to indicate the level corresponding to users with lower risk tolerance, and the second level may be used to indicate the level corresponding to users with higher risk tolerance. The preset similarity threshold is a preset similarity threshold between the first similarity and the second similarity, which may be any value and is not specifically limited here.

[0074] Specifically, if the pet claim risk level is the first level, when the first similarity and the second similarity both exceed the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet; when the first similarity or the second similarity does not exceed the preset similarity threshold, it is determined that the identity of the pet to be detected is different from that of the target pet.

[0075] Exemplarily, assuming that the pet claim risk level is the first level, the preset similarity threshold corresponding to the first similarity is 0.8, and the preset similarity threshold corresponding to the second similarity is 0.9. If the first similarity is 0.9 and the second similarity is 0.95, and both the first similarity and the second similarity exceed the corresponding preset similarity thresholds, it is determined that the identity of the pet to be detected is the same as that of the target pet; if the first similarity is 0.9 and the second similarity is 0.85, and the second similarity does not exceed the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is different from that of the target pet.

[0076] Specifically, if the pet claim risk level is the second level, when the first similarity or the second similarity exceeds the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet; when both the first similarity and the second similarity do not exceed the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is different from that of the target pet.

[0077] Exemplarily, assuming that the pet claim risk level is the second level, the preset similarity threshold corresponding to the first similarity is 0.8, and the preset similarity threshold corresponding to the second similarity is 0.9. If the first similarity is 0.8 and the second similarity is 0.95, and the second similarity exceeds the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet; if the first similarity is 0.8 and the second similarity is 0.85, and both the first similarity and the second similarity do not exceed the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is different from that of the target pet.

[0078] In some embodiments, the pet claim risk level is obtained, and a first similarity threshold and a second similarity threshold are determined according to the pet claim risk level; when the first similarity exceeds the first similarity threshold and the second similarity exceeds the second similarity threshold, it is determined that the identity of the pet to be detected and the target pet are the same. In this way, the identity information of the pet to be detected and the target pet can be accurately determined, thereby avoiding pet image fraud, strictly controlling the risk of pet insurance, and greatly reducing the risk of insurance fraud.

[0079] Among them, the first similarity threshold is a similarity threshold corresponding to the first similarity determined according to the pet claim risk level; the second similarity threshold is a similarity threshold corresponding to the second similarity determined according to the pet claim risk level.

[0080] Specifically, a first similarity threshold and a second similarity threshold are determined according to the pet claim risk level; when the first similarity exceeds the first similarity threshold and the second similarity exceeds the second similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet; when the first similarity does not exceed the first similarity threshold or the second similarity does not exceed the second similarity threshold, it is determined that the identity of the pet to be detected is not the same as that of the target pet.

[0081] Exemplarily, the first similarity threshold and the second similarity threshold can be determined as 0.9 and 0.8 respectively according to the pet claim risk level. If the first similarity is 0.95 and the second similarity is 0.85, the first similarity exceeds the first similarity threshold and the second similarity exceeds the second similarity threshold, then it is determined that the identity of the pet to be detected is the same as that of the target pet; if the first similarity is 0.95 and the second similarity is 0.75, the first similarity exceeds the first similarity threshold but the second similarity does not exceed the second similarity threshold, then it is determined that the identity of the pet to be detected is different from that of the target pet.

[0082] See also Figure 3 , Figure 3 It is a schematic block diagram of a pet identity verification device provided in one embodiment of the present application. The pet identity verification device can be configured in a server to execute the aforementioned pet identity verification method.

[0083] like Figure 3 As shown, the pet identity verification device 200 includes: an audio and video acquisition module 201, a voiceprint feature extraction module 202, an appearance feature extraction module 203, an identity feature acquisition module 204, a feature comparison module 205 and an identity recognition module 206.

[0084] The audio and video acquisition module 201 is used to acquire audio and video information of the pet to be detected, wherein the audio and video information includes a pet video and a pet audio corresponding to the pet to be detected;

[0085] A voiceprint feature extraction module 202 is used to extract and identify the pet audio to obtain the pet voiceprint feature;

[0086] Appearance feature extraction module 203, used for extracting features and performing recognition processing on the pet video to obtain pet appearance features;

[0087] An identity feature acquisition module 204 is used to acquire identity feature information of a target pet, wherein the identity feature information includes a target voiceprint feature and a target appearance feature;

[0088] A feature comparison module 205 is used to determine the similarity between the pet voiceprint feature and the target voiceprint feature as a first similarity, and to determine the similarity between the pet appearance feature and the target appearance feature as a second similarity;

[0089] The identity recognition module 206 is used to generate an identity verification result of the pet to be detected and the target pet according to the first similarity and the second similarity.

[0090] In some embodiments, the voiceprint feature extraction module 202 is also used to crop and denoise the pet audio to obtain optimized pet audio; based on the pet call feature extraction model, feature extraction is performed on the optimized pet audio to obtain pet voiceprint features.

[0091] In some embodiments, the appearance feature extraction module 203 is also used to perform image detection processing on the pet video to obtain a pet multi-dimensional image; perform target detection on the pet multi-dimensional image to obtain pet object information corresponding to the pet multi-dimensional image; based on a pet appearance feature extraction model, perform feature extraction on the pet object information corresponding to the pet multi-dimensional image to obtain pet appearance features.

[0092] In some embodiments, the appearance feature extraction module 203 is also used to perform feature extraction on the pet object information corresponding to the pet's front face image to obtain the pet's front face features; perform feature extraction on the pet object information corresponding to the pet's side view image to obtain the pet's side view features; perform feature extraction on the pet object information corresponding to the pet's overhead view image to obtain the pet's overhead view features; generate the pet's appearance features based on the pet's front face features, the pet's side view features and the pet's overhead view features.

[0093] In some embodiments, the appearance feature extraction module 203 is also used to obtain image information corresponding to the pet's front face image, pet's side image and pet's overhead view image, wherein the image information includes image clarity and target object ratio; determine the weight ratio corresponding to the pet's front face image, pet's side image and pet's overhead view image according to the image clarity and the target object ratio; generate the pet's appearance features according to the pet's front face features, the pet's side features and the pet's overhead view features and their corresponding weight ratios.

[0094] In some embodiments, the identity recognition module 206 is also used to obtain the pet claim risk level; if the pet claim risk level is the first level, when the first similarity and the second similarity both exceed the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet; if the pet claim risk level is the second level, when the first similarity or the second similarity exceeds the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet.

[0095] In some embodiments, the identity recognition module 206 is also used to obtain the pet claim risk level, and determine a first similarity threshold and a second similarity threshold based on the pet claim risk level; when the first similarity exceeds the first similarity threshold and the second similarity exceeds the second similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet.

[0096] It should be noted that those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices, modules and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0097] The method and apparatus of the present application can be used in many general or special computing system environments or configurations, such as personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer terminal devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices.

[0098] Exemplarily, the above method and apparatus may be implemented in the form of a computer program. The computer program may be implemented in the form of a computer program. Figure 4 Runs on the computer device shown.

[0099] See also Figure 4 , Figure 4 1 is a schematic diagram of a computer device provided in an embodiment of the present application. The computer device may be a server.

[0100] like Figure 4 As shown, the computer device includes a processor, a memory and a network interface connected through a system bus, wherein the memory may include a volatile storage medium, a non-volatile storage medium and an internal memory.

[0101] The non-volatile storage medium can store an operating system and a computer program. The computer program includes program instructions, and when the program instructions are executed, the processor can execute any pet identity verification method.

[0102] The processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.

[0103] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor, the processor can execute any pet identity verification method.

[0104] The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art can understand that the structure of the computer device is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0105] It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0106] Among them, in some embodiments, the processor is used to run a computer program stored in the memory to implement the following steps: obtain audio and video information of the pet to be detected, the audio and video information including the pet video and pet audio corresponding to the pet to be detected; perform feature extraction and recognition processing on the pet audio to obtain the pet voiceprint feature; perform feature extraction and recognition processing on the pet video to obtain the pet appearance feature; obtain the identity feature information of the target pet, the identity feature information including the target voiceprint feature and the target appearance feature; determine the similarity between the pet voiceprint feature and the target voiceprint feature as a first similarity, and determine the similarity between the pet appearance feature and the target appearance feature as a second similarity; generate the identity verification result of the pet to be detected and the target pet based on the first similarity and the second similarity.

[0107] In some embodiments, the processor is also used to crop and denoise the pet audio to obtain optimized pet audio; based on a pet call feature extraction model, feature extraction is performed on the optimized pet audio to obtain pet voiceprint features.

[0108] In some embodiments, the processor is also used to perform image detection processing on the pet video to obtain a pet multi-dimensional image; perform target detection on the pet multi-dimensional image to obtain pet object information corresponding to the pet multi-dimensional image; and perform feature extraction on the pet object information corresponding to the pet multi-dimensional image based on a pet appearance feature extraction model to obtain pet appearance features.

[0109] In some embodiments, the processor is further used to perform feature extraction on the pet object information corresponding to the pet's front face image to obtain the pet's front face features; perform feature extraction on the pet object information corresponding to the pet's side view image to obtain the pet's side view features; perform feature extraction on the pet object information corresponding to the pet's overhead view image to obtain the pet's overhead view features; generate the pet's appearance features based on the pet's front face features, the pet's side view features and the pet's overhead view features.

[0110] In some embodiments, the processor is also used to obtain image information corresponding to the pet's front face image, pet's side view image, and pet's overhead view image, the image information including image clarity and target object ratio; determine the weight ratio corresponding to the pet's front face image, pet's side view image, and pet's overhead view image according to the image clarity and the target object ratio; generate the pet's appearance features according to the pet's front face features, the pet's side view features, the pet's overhead view features and their corresponding weight ratios.

[0111] In some embodiments, the processor is also used to obtain the pet claim risk level; if the pet claim risk level is the first level, when the first similarity and the second similarity both exceed the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet; if the pet claim risk level is the second level, when the first similarity or the second similarity exceeds the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet.

[0112] In some embodiments, the processor is also used to obtain a pet claim risk level, and determine a first similarity threshold and a second similarity threshold based on the pet claim risk level; when the first similarity exceeds the first similarity threshold and the second similarity exceeds the second similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet.

[0113] An embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. The computer program includes program instructions, and when the program instructions are executed, any one of the pet identity verification methods provided in the embodiments of the present application is implemented.

[0114] The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiment, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc., equipped on the computer device.

[0115] Furthermore, the computer-readable storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the blockchain node, etc.

[0116] The present invention refers to a new application model of computer technologies such as storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. of the blockchain language model. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the blockchain underlying platform, platform product service layer, and application service layer.

[0117] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A pet identity verification method, It is characterized in that The method comprises: Acquire audio and video information of the pet to be detected, wherein the audio and video information includes a pet video and a pet audio corresponding to the pet to be detected; Extracting and identifying the pet audio to obtain the pet voiceprint feature; Performing feature extraction and recognition processing on the pet video to obtain the pet's appearance features; Acquire identity feature information of a target pet, wherein the identity feature information includes a target voiceprint feature and a target appearance feature; Determine the similarity between the pet's voiceprint feature and the target voiceprint feature as a first similarity, and determine the similarity between the pet's appearance feature and the target appearance feature as a second similarity; An identity verification result of the pet to be detected and the target pet is generated according to the first similarity and the second similarity.

2. The method according to claim 1, It is characterized in that The extracting and identifying the pet audio to obtain the pet voiceprint feature includes: Performing cropping and denoising on the pet audio to obtain optimized pet audio; Based on the pet call feature extraction model, feature extraction is performed on the optimized pet audio to obtain the pet voiceprint feature.

3. The method according to claim 1, It is characterized in that The feature extraction and recognition processing of the pet video to obtain the pet appearance features includes: Performing image detection processing on the pet video to obtain a multi-dimensional image of the pet; Performing target detection on the pet multi-dimensional image to obtain pet object information corresponding to the pet multi-dimensional image; Based on the pet appearance feature extraction model, feature extraction is performed on the pet object information corresponding to the pet multi-dimensional image to obtain the pet appearance features.

4. The method according to claim 3, It is characterized in that The pet multi-dimensional image includes a pet front face image, a pet side image and a pet top view image. The pet appearance feature extraction model is based on which the pet object information corresponding to the pet multi-dimensional image is extracted to obtain the pet appearance features, including: Extracting features of the pet object information corresponding to the pet front face image to obtain pet front face features; Extracting features of the pet object information corresponding to the pet side image to obtain pet side features; Extracting features of the pet object information corresponding to the pet bird's-eye view image to obtain pet bird's-eye view features; The pet's appearance features are generated based on the pet's frontal facial features, the pet's sideways features and the pet's overhead features.

5. The method according to claim 4, It is characterized in that The generating of the pet appearance features according to the pet frontal face features, the pet side body features and the pet top view features includes: Obtaining image information corresponding to the pet's front face image, the pet's side image, and the pet's top view image, wherein the image information includes image clarity and a target object ratio; Determine the weight ratios corresponding to the pet front face image, the pet side image, and the pet top view image according to the image clarity and the target object proportion; The pet appearance features are generated according to the pet's frontal facial features, the pet's sideways features, the pet's overhead features and their corresponding weight ratios.

6. The method according to claim 1, It is characterized in that The step of generating an identity verification result between the pet to be detected and the target pet according to the first similarity and the second similarity includes: Get pet claims risk level; If the pet claim risk level is the first level, when both the first similarity and the second similarity exceed the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet; If the pet claim risk level is the second level, when the first similarity or the second similarity exceeds the corresponding preset similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet.

7. The method according to claim 1, It is characterized in that The step of generating an identity verification result between the pet to be detected and the target pet according to the first similarity and the second similarity includes: Obtaining a pet claim risk level, and determining a first similarity threshold and a second similarity threshold according to the pet claim risk level; When the first similarity exceeds the first similarity threshold and the second similarity exceeds the second similarity threshold, it is determined that the identity of the pet to be detected is the same as that of the target pet.

8. A pet identity verification device, It is characterized in that The pet identity verification device comprises: An audio and video acquisition module, used to acquire audio and video information of the pet to be detected, wherein the audio and video information includes a pet video and a pet audio corresponding to the pet to be detected; A voiceprint feature extraction module, used to extract and identify the pet audio to obtain the pet voiceprint feature; An appearance feature extraction module is used to extract and identify the pet video to obtain the pet's appearance features; An identity feature acquisition module, used to acquire identity feature information of a target pet, wherein the identity feature information includes a target voiceprint feature and a target appearance feature; A feature comparison module, configured to determine the similarity between the pet's voiceprint feature and the target voiceprint feature as a first similarity, and to determine the similarity between the pet's appearance feature and the target appearance feature as a second similarity; An identity recognition module is used to generate an identity verification result of the pet to be detected and the target pet according to the first similarity and the second similarity.

9. A computer device, It is characterized in that The computer device includes a memory and a processor; The memory is used to store computer programs; The processor is configured to execute the computer program and implement, when executing the computer program: A pet identity verification method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor implements the pet identity verification method according to any one of claims 1 to 7.

Images