A method, apparatus, electronic device, and storage medium for identifying living organisms

Abstract

This application provides a method, apparatus, electronic device, and storage medium for determining liveness. The method includes: acquiring a video of a user reading text to be judged; determining that the lip shape category corresponding to the text is the same in different dialects; acquiring lip shape coordinate information corresponding to each video frame in the video of the text to be judged; inputting all the lip shape coordinate information sequentially into a lip shape detection model to obtain the user's lip shape category in the video; the lip shape detection model is trained using lip shape coordinate information samples and corresponding lip shape categories; determining whether the user's lip shape category is the same as the lip shape category corresponding to the text to be judged; if the user's lip shape category is the same as the lip shape category corresponding to the text to be judged, then determining that the user is a live person. This method improves the accuracy of determining whether a user is a live person.

Description

Technical Field

[0001] This invention relates to the field of biometrics, and more specifically, to a method, apparatus, electronic device, and storage medium for identifying a living person. Background Technology

[0002] With the development of mobile internet, liveness detection has received increasing attention in fields such as finance, insurance, banking, and securities. Currently, commonly used liveness detection technologies include reading-based liveness recognition, which uses audio recordings of digital pronunciation as the basis for identification.

[0003] However, liveness detection for reading numbers only uses the audio of the digit's pronunciation as the basis for judgment. But different numbers may sound similar in different dialects, making the audio of the digit's pronunciation an inaccurate basis for judgment. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a method, device, electronic device and storage medium for determining whether a user is a living person, thereby improving the accuracy of determining whether a user is a living person.

[0005] In a first aspect, embodiments of this application provide a method for identifying a living organism, the method comprising:

[0006] Acquire videos of users reading and judging text; determine if the lip movements corresponding to the text are the same in different dialects;

[0007] Obtain the mouth shape coordinates for each video frame in the video used to determine the text.

[0008] All mouth shape coordinates are sequentially input into the mouth shape detection model to obtain the mouth shape category of the user in the video; the mouth shape detection model is trained using mouth shape coordinate information samples and corresponding mouth shape categories.

[0009] Determine whether the user's lip shape category matches the lip shape category corresponding to the judgment text;

[0010] If the user's mouth shape category matches the mouth shape category corresponding to the judgment text, then the user is determined to be alive.

[0011] In one possible implementation, the mouth shape categories include a mouth-opening-then-closing category, a mouth-closing-then-opening category, and a pouting category; the method further includes:

[0012] Select one character from each of the preset text categories (open mouth then close mouth, close mouth then open mouth, pout) and combine them to form the judgment text for the user.

[0013] In one possible implementation, obtaining the lip shape coordinate information corresponding to each video frame in the video of the determined text includes:

[0014] The initial lip shape coordinates of each video frame are detected using a facial landmark detection model.

[0015] The initial mouth shape coordinate information is filtered using the Kalman filter algorithm to obtain the target mouth shape coordinate information.

[0016] In one possible implementation, before determining whether the user's lip shape category is the same as the lip shape category corresponding to the judgment text, the method further includes:

[0017] Obtain audio of the user reading the text to determine its accuracy;

[0018] Convert audio into text to obtain the text that the user is reading from the audio.

[0019] After determining that the user's lip shape category matches the lip shape category corresponding to the judgment text, the method further includes:

[0020] Determine whether the text being read by the user is the same as the text being judged;

[0021] If the text being read by the user is the same as the text being judged, then the user is determined to be a live entity.

[0022] In one possible implementation, the audio is converted into text to obtain the text read by the user in the audio, including:

[0023] Audio is converted into text using automatic speech recognition technology, resulting in the first text that the user reads from the audio.

[0024] In one possible implementation, converting audio into text to obtain the text read by the user in the audio further includes:

[0025] After adding Gaussian noise to the audio, the audio with added Gaussian noise is converted into text using automatic speech recognition technology, resulting in the second reading text that the user reads from the audio.

[0026] In one possible implementation, determining whether the text being read by the user is the same as the text to be determined includes:

[0027] If the first or second reading text is the same as the judgment text, then the reading text read by the user is the same as the judgment text.

[0028] Secondly, embodiments of this application also provide a living body detection device, which includes:

[0029] The acquisition module is used to acquire videos of users reading and judging text; the lip shape categories corresponding to the judged text are the same in different dialects.

[0030] The acquisition module is also used to acquire the mouth shape coordinate information corresponding to each video frame in the video of the text to be judged;

[0031] The input module is used to sequentially input all mouth shape coordinate information into the mouth shape detection model to obtain the mouth shape category of the user in the video; the mouth shape detection model is trained by mouth shape coordinate information samples and corresponding mouth shape categories;

[0032] The judgment module is used to determine whether the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text; if the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text, then the user is determined to be alive.

[0033] In one possible implementation, the mouth shape categories include a mouth-opening-then-closing category, a mouth-closing-then-opening category, and a pouting category; the device also includes a selection module;

[0034] The selection module is used to select one character from the preset characters corresponding to the categories of "open mouth then close mouth", "close mouth then open mouth", and "pouting mouth" and combine them as the judgment text for the user.

[0035] In one possible implementation, the acquisition module is specifically used to detect the initial lip shape coordinate information corresponding to each video frame in the video through a facial key point detection model; and to filter the initial lip shape coordinate information through a Kalman filter algorithm to obtain the target lip shape coordinate information.

[0036] In one possible implementation, the device further includes: a conversion module;

[0037] The acquisition module is also used to acquire the audio of the user reading the text for judgment;

[0038] The conversion module is used to convert audio into text, obtaining the text that the user reads from the audio.

[0039] The judgment module is also used to determine whether the text read by the user is the same as the judgment text; if the text read by the user is the same as the judgment text, then the user is determined to be a live object.

[0040] In one possible implementation, the conversion module is specifically used to convert audio into text using automatic speech recognition technology, thereby obtaining the first reading text read by the user in the audio.

[0041] In one possible implementation, the conversion module is specifically used to add Gaussian noise to the audio and then convert the audio with added Gaussian noise into text using automatic speech recognition technology, thereby obtaining the second reading text read by the user in the audio.

[0042] In one possible implementation, the judgment module is specifically configured to determine if the first reading text or the second reading text is the same as the judgment text, and therefore the reading text read by the user is the same as the judgment text.

[0043] Thirdly, embodiments of this application also provide an electronic device, including: a processor, a storage medium, and a bus. The storage medium stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of any of the methods for determining a living person in the first aspect.

[0044] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of any of the methods for determining a living being in the first aspect.

[0045] This application provides a method, apparatus, electronic device, and storage medium for determining liveness. The method includes: acquiring a video of a user reading text to be judged; determining that the lip shape category corresponding to the text is the same in different dialects; acquiring lip shape coordinate information corresponding to each video frame in the video of the text to be judged; inputting all the lip shape coordinate information sequentially into a lip shape detection model to obtain the user's lip shape category in the video; the lip shape detection model is trained using lip shape coordinate information samples and corresponding lip shape categories; determining whether the user's lip shape category is the same as the lip shape category corresponding to the text to be judged; if the user's lip shape category is the same as the lip shape category corresponding to the text to be judged, then the user is determined to be live. This application improves the accuracy of determining whether a user is live by judging whether the user's lip shape category in a video of a user reading text with the same lip shape category in different dialects is the same as the lip shape category corresponding to the text to be judged. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 A flowchart of a method for determining a living organism provided in an embodiment of this application is shown;

[0048] Figure 2 A flowchart of another method for determining a living organism provided in an embodiment of this application is shown;

[0049] Figure 3A flowchart of another method for determining a living organism provided in an embodiment of this application is shown;

[0050] Figure 4 This illustration shows a schematic diagram of a living organism detection device provided in an embodiment of this application;

[0051] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0053] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0054] To enable those skilled in the art to utilize the content of this application, and in conjunction with the specific application scenario of "the field of biometric technology," the following embodiments are provided. For those skilled in the art, the general principles defined herein can be applied to other embodiments and application scenarios without departing from the spirit and scope of this application. Although this application is primarily described within the "field of biometric technology," it should be understood that this is merely an exemplary embodiment.

[0055] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0056] The following is a detailed description of a method for identifying living organisms provided in the embodiments of this application.

[0057] Reference Figure 1 The diagram shown is a flowchart illustrating a method for identifying a living organism according to an embodiment of this application. The specific execution process of this method is as follows:

[0058] S101. Obtain a video of the user reading the text to determine the identity; determine that the lip shape category corresponding to the text is the same in different dialects.

[0059] S102. Obtain the mouth shape coordinate information corresponding to each video frame in the video of the determined text.

[0060] S103. Input all mouth shape coordinates into the mouth shape detection model in sequence to obtain the mouth shape category of the user in the video.

[0061] S104. Determine whether the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text.

[0062] S105. If the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text, then the user is determined to be a live object.

[0063] This application provides a method for determining liveness, which includes: acquiring a video of a user reading a judgment text; determining that the judgment text corresponds to the same lip shape category in different dialects; acquiring lip shape coordinate information corresponding to each video frame in the video of the judgment text; inputting all lip shape coordinate information sequentially into a lip shape detection model to obtain the user's lip shape category in the video; the lip shape detection model is trained using lip shape coordinate information samples and corresponding lip shape categories; determining whether the user's lip shape category is the same as the lip shape category corresponding to the judgment text; if the user's lip shape category is the same as the lip shape category corresponding to the judgment text, then the user is determined to be live. This application improves the accuracy of determining whether a user is live by judging whether the user's lip shape category in a video of a judgment text with the same lip shape category in different dialects is the same as the lip shape category corresponding to the judgment text.

[0064] The exemplary steps of the embodiments of this application are described below:

[0065] S101. Obtain a video of the user reading the judgment text.

[0066] In this application's implementation, the judgment text includes multiple characters. The user reads the characters in the judgment text sequentially and obtains a video of reading the judgment text. The judgment text refers to the text used for liveness detection of the user, and the corresponding mouth shape category is the same across different dialects. The mouth shape categories include the open-mouth-then-closed-mouth category, the closed-mouth-then-open-mouth category, and the pouting category. The mouth shape category refers to the category of mouth shape changes when a person reads text. According to the vertical opening and closing relationship of the mouth shape, it can be divided into open-mouth-then-closed, closed-mouth-then-open, and left-right opening and closing. The mouth shape when a person reads all text falls within these three vertical opening and closing relationships. Since an open mouth means opening the mouth and a closed mouth means closing the mouth, open-mouth-then-closed corresponds to the open-mouth-then-closed-mouth category, closed-mouth-then-open corresponds to the closed-mouth-then-open-mouth category, and left-right opening and closing corresponds to the pouting category.

[0067] Here, because the mouth shapes for certain characters differ in different dialects, in order to improve the accuracy of live identification, the mouth shape category corresponding to the character must be the same in different dialects.

[0068] In this application embodiment, one character is selected from the preset characters corresponding to the categories of opening the mouth before closing the mouth, closing the mouth before opening the mouth, and pouting, and then combined to serve as the user's judgment character.

[0069] Here, each mouth shape category corresponds to at least one preset character, and each preset character corresponds to the same mouth shape category in different dialects, so as to ensure that the judgment character selected from the preset characters corresponds to the same mouth shape category in different dialects.

[0070] The preset text for the "open mouth then close mouth" category can be "ah," "tooth," "big," etc.; the preset text for the "close mouth then open mouth" category can be "closed," "skin," "rice," etc.; and the preset text for the "pouting mouth" category can be "du," "book," "house," etc. For example, if the preset text for the "open mouth then close mouth" category is "ah," the preset text for the "closed mouth then open mouth" category is "closed," and the preset text for the "pouting mouth" category is "du," then the combination "ah," "closed," and "du" will be used as the user's judgment text. The user will then read "ah," "closed," and "du" in sequence.

[0071] S102. Obtain the mouth shape coordinate information corresponding to each video frame in the video of the determined text.

[0072] In this embodiment, lip shape coordinate information refers to the coordinate information of the mouth region in the video. If the coordinates of the mouth region are represented by the coordinates of m points within the mouth region, including the x-coordinate and y-coordinate of each point, then the lip shape coordinate information for each frame is represented by an m*2 array. If the video has n video frames, then the lip shape coordinate information for the entire video segment is represented by an n*m*2 array.

[0073] The initial lip shape coordinates of each video frame in the video are detected by a facial landmark detection model; the initial lip shape coordinates are then filtered by a Kalman filter algorithm to obtain the target lip shape coordinates.

[0074] In this application's embodiments, the facial landmark detection model can be the Meidiapipe facial landmark detection model, or other models capable of extracting lip coordinate information. Since the extracted lip coordinate information may contain errors due to factors such as movement of the shooting device during video recording, this application uses a Kalman filter algorithm to filter the initial lip coordinate information for each frame after the first video frame, thereby reducing the error.

[0075] Furthermore, the initial mouth shape coordinate information is filtered using a Kalman filter algorithm to obtain the target mouth shape coordinate information, specifically including the following steps:

[0076] I. Determine the initial mouth shape coordinates corresponding to the first frame as the target mouth shape coordinates.

[0077] II. For each frame other than the first frame, based on the initial mouth shape coordinate information corresponding to the frame and the target mouth shape coordinate information corresponding to the previous frame, the initial mouth shape coordinate information corresponding to the frame is filtered by the following Kalman filter algorithm to obtain the target mouth shape coordinate information corresponding to the frame.

[0078] x k =f(x) k-1 )+A(y k -Cf(x k-1 ));

[0079] Where, x k f(x) represents the target mouth shape coordinates corresponding to this frame. k-1 ) represents the target mouth shape coordinates corresponding to the previous frame in this frame, A is the Kalman gain, and y k Here, C represents the initial mouth shape coordinates corresponding to this frame, and C is the weighting coefficient between the Kalman prediction and the measurement.

[0080] S103. Input all mouth shape coordinates into the mouth shape detection model in sequence to obtain the mouth shape category of the user in the video.

[0081] In this embodiment, a Long Short-Term Memory (LSTM) model is used as the lip shape detection model. This lip shape detection model is trained using lip shape coordinate information samples and corresponding lip shape categories.

[0082] Since mouth movements themselves constitute a time-series data set, LSTM (Long Short-Term Memory) networks are naturally well-suited for time-series data. Extensive theoretical and empirical evidence demonstrates that LSTM performs significantly better than other traditional machine learning methods such as SVM and Random Forest on time-series data. This paper creatively proposes using grid search to determine the parameters when defining the LSTM network structure. This approach allows for the most reasonable and optimal construction of an LSTM model that matches the data, further improving the model's accuracy.

[0083] Here, for LSTM, the three core parameters are the number of hidden neurons, the number of recurrent layers, and the activation functions of the hidden neurons. These parameters are determined through a grid search, i.e., by constructing different structures and comparing the performance of different networks on the validation set to decide which parameters to use. The grid search ranges for the three parameters are: number of hidden neurons (40–100), number of recurrent layers (4–8), and activation functions of the hidden neurons (leak_relu, relu, softmax). After grid search validation, the final model structure parameters are: 72 hidden neurons, 5 recurrent layers, and the activation function of the hidden neurons is leak_relu. Furthermore, L2 regularization is used to constrain the parameters and improve the model's generalization ability; the L2 parameter is set to 1e-6.

[0084] Furthermore, the embodiments of this application are not limited to using the LSTM model as the mouth shape detection model; other models that can achieve the same effect can also be used as the mouth shape detection model.

[0085] S104. Determine whether the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text.

[0086] In this embodiment, the user's lip shape category corresponds to the order of the judgment text. For example, if the user's lip shape categories are a, b, and c, and the judgment texts are A, B, and C, then it is determined whether 'a' and the lip shape category corresponding to judgment text A are the same, whether 'b' and the lip shape category corresponding to judgment text B are the same, and whether 'c' and the lip shape category corresponding to judgment text C are the same. When the number of identical lip shapes exceeds a preset value, the user's lip shape category is considered to be the same as the lip shape category corresponding to the judgment text.

[0087] S105. If the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text, then the user is determined to be a live object.

[0088] This application provides a method for determining liveness, which includes: acquiring a video of a user reading text to be judged; determining that the lip shape category corresponding to the text is the same in different dialects; acquiring lip shape coordinate information corresponding to each video frame in the video of the text to be judged; inputting all the lip shape coordinate information sequentially into a lip shape detection model to obtain the user's lip shape category in the video; the lip shape detection model is trained using lip shape coordinate information samples and corresponding lip shape categories; determining whether the user's lip shape category is the same as the lip shape category corresponding to the text to be judged; if the user's lip shape category is the same as the lip shape category corresponding to the text to be judged, then the user is determined to be live. This application improves the accuracy of determining whether a user is live by judging whether the user's lip shape category in the video of the text to be judged, which has the same lip shape category in different dialects, is the same as the lip shape category corresponding to the text to be judged.

[0089] Reference Figure 2 The diagram shown is a flowchart illustrating another method for determining a living organism provided in this application. The following steps are related to... Figure 1 The steps that are the same in the previous steps will not be described again. The following describes the exemplary steps of the embodiments of this application:

[0090] S201. Obtain the video of the user reading the judgment text. And obtain the audio of the user reading the judgment text.

[0091] The character being judged has the same mouth shape and pronunciation in different dialects, while the pronunciation of other characters is different from that of the character being judged in different dialects.

[0092] S202. Obtain the lip shape coordinate information corresponding to each video frame in the video of the determined text, and convert the audio into text to obtain the reading text read by the user in the audio.

[0093] Converting audio into text to obtain the reading text described in the audio for the user to read includes several methods.

[0094] Optionally, the audio is converted into text using automatic speech recognition technology to obtain the first reading text read by the user in the audio, and the first reading text is used as the reading text read by the user.

[0095] Here, in this embodiment of the application, audio can be converted into text using ASR (automatic speech recognition) technology to obtain the first reading text read by the user in the audio, and the first reading text is used as the reading text read by the user.

[0096] Optionally, after adding Gaussian noise to the audio, the audio with added Gaussian noise is converted into text using automatic speech recognition technology to obtain the second reading text read by the user in the audio, and the second reading text is used as the reading text read by the user.

[0097] Optionally, both the first and second reading texts mentioned above can be used as the reading texts for the user.

[0098] S203. Input all mouth shape coordinates into the mouth shape detection model in sequence to obtain the mouth shape category of the user in the video.

[0099] S204. Determine whether the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text.

[0100] S205. If the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text, then determine whether the reading text read by the user is the same as the judgment text; if the reading text read by the user is the same as the judgment text, then determine that the user is a living person.

[0101] The system determines whether the text being read by the user matches the text being judged. If the text matches, the user is confirmed to be a live user, and this can be achieved through various methods, including:

[0102] Optionally, if the first reading text is taken as the reading text read by the user, then it is determined whether the first reading text is the same as the judgment text; if the first reading text is the same as the judgment text, then the user is determined to be a living person.

[0103] Optionally, if the second reading text is used as the reading text read by the user, then it is determined whether the second reading text is the same as the judgment text; if the second reading text is the same as the judgment text, then the user is determined to be a live entity.

[0104] Optionally, if both the first and second reading texts are taken as the reading texts read by the user, then it is determined whether the first or second reading text is the same as the determined text. If the first or second reading text is the same as the determined text, then the reading text read by the user is the same as the determined text.

[0105] Optionally, if both the first and second reading texts are taken as the reading texts read by the user, then it is determined whether both the first and second reading texts are the same as the determined text. If both the first and second reading texts are the same as the determined text, then the reading texts read by the user are the same as the determined text.

[0106] This application provides another method for determining liveness, which includes: acquiring a video of a user reading judgment text; acquiring audio of the user reading the judgment text; acquiring lip shape coordinate information corresponding to each video frame in the video of the judgment text, and converting the audio into text to obtain the text read by the user in the audio; inputting all the lip shape coordinate information sequentially into a lip shape detection model to obtain the lip shape category of the user in the video; determining whether the user's lip shape category is the same as the lip shape category corresponding to the judgment text; if the user's lip shape category is the same as the lip shape category corresponding to the judgment text, then determining whether the text read by the user is the same as the judgment text; if the text read by the user is the same as the judgment text, then determining that the user is a live person. This method can determine whether a user is a live person, improving the accuracy of user liveness determination.

[0107] Reference Figure 3 The diagram shown is a flowchart illustrating another method for determining a living organism provided in this application. The following steps are related to... Figure 1 or Figure 2 The steps that are the same in the previous steps will not be described again. The following describes the exemplary steps of the embodiments of this application:

[0108] S301. Obtain the video of the user reading the judgment text. And obtain the audio of the user reading the judgment text.

[0109] S302. Obtain the lip shape coordinate information corresponding to each video frame in the video of the determined text, and convert the audio into text to obtain the reading text read by the user in the audio.

[0110] S303. Input all mouth shape coordinates into the mouth shape detection model in sequence to obtain the mouth shape category of the user in the video.

[0111] S304. Determine whether the user's lip shape category is the same as the lip shape category corresponding to the judgment text; and determine whether the reading text read by the user is the same as the judgment text.

[0112] S305. If the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text, or if the user's reading text is the same as the judgment text, then the user is determined to be a living person.

[0113] This application provides another method for determining whether a user is alive. By using this method, it is possible to determine whether a user is alive, thereby improving the accuracy of determining whether a user is alive.

[0114] Based on the same inventive concept, this application also provides a living body identification device corresponding to the living body identification method. Since the principle of the device in this application is similar to the living body identification method described above in this application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0115] Reference Figure 4 The diagram shown is a schematic of a liveness detection device provided in an embodiment of this application. The liveness detection device includes:

[0116] The acquisition module 401 is used to acquire a video of the user reading the judgment text; the judgment text corresponds to the same lip shape category in different dialects.

[0117] The acquisition module 401 is also used to acquire the mouth shape coordinate information corresponding to each video frame in the video of the determined text;

[0118] The input module 402 is used to input all mouth shape coordinate information into the mouth shape detection model in sequence to obtain the mouth shape category of the user in the video; the mouth shape detection model is trained by mouth shape coordinate information samples and corresponding mouth shape categories;

[0119] The judgment module 403 is used to determine whether the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text; if the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text, then the user is determined to be alive.

[0120] In one possible implementation, the mouth shape categories include a mouth-opening-then-closing category, a mouth-closing-then-opening category, and a pouting category; the device also includes: a selection module 404;

[0121] The selection module 404 is used to select one character from the preset characters corresponding to the categories of "open mouth then close mouth", "close mouth then open mouth", and "pouting mouth" respectively, and combine them as the judgment text for the user.

[0122] In one possible implementation, the acquisition module 401 is specifically used to detect the initial mouth shape coordinate information corresponding to each video frame in the video through a face key point detection model; and to filter the initial mouth shape coordinate information through a Kalman filter algorithm to obtain the target mouth shape coordinate information.

[0123] In one possible implementation, the device further includes: a conversion module 405;

[0124] The acquisition module 401 is also used to acquire the audio of the user reading the judgment text;

[0125] The conversion module 405 is used to convert audio into text, obtaining the reading text that the user reads in the audio.

[0126] The judgment module 403 is also used to determine whether the reading text read by the user is the same as the judgment text; if the reading text read by the user is the same as the judgment text, then the user is determined to be a living person.

[0127] In one possible implementation, the conversion module 405 is specifically used to convert audio into text using automatic speech recognition technology to obtain the first reading text read by the user in the audio.

[0128] In one possible implementation, the conversion module 405 is specifically used to add Gaussian noise to the audio and then convert the audio with added Gaussian noise into text using automatic speech recognition technology to obtain the second reading text read by the user in the audio.

[0129] In one possible implementation, the judgment module 403 is specifically used to determine if the first reading text or the second reading text is the same as the judgment text, then the reading text read by the user is the same as the judgment text.

[0130] This application provides a liveness detection device, comprising: an acquisition module 401 for acquiring a video of a user reading judgment text; the judgment text corresponds to the same lip shape category in different dialects; the acquisition module 401 is further used to acquire lip shape coordinate information corresponding to each video frame in the video of the judgment text; an input module 402 for sequentially inputting all lip shape coordinate information into a lip shape detection model to obtain the user's lip shape category in the video; the lip shape detection model is trained using lip shape coordinate information samples and corresponding lip shape categories; and a judgment module 403 for judging whether the user's lip shape category is the same as the lip shape category corresponding to the judgment text; if the user's lip shape category is the same as the lip shape category corresponding to the judgment text, then the user is determined to be a live person. This application improves the accuracy of judging whether a user is a live person by judging whether the user's lip shape category in a video of a user reading judgment text with the same lip shape category in different dialects is the same as the lip shape category corresponding to the judgment text.

[0131] like Figure 5 As shown in the embodiment of this application, an electronic device 500 includes a processor 501, a memory 502, and a bus. The memory 502 stores machine-readable instructions that can be executed by the processor 501. When the electronic device is running, the processor 501 and the memory 502 communicate through the bus. The processor 501 executes the machine-readable instructions to perform the steps of the above-described method for determining a living person.

[0132] Specifically, the memory 502 and processor 501 mentioned above can be general-purpose memory and processor, without any specific limitations. When the processor 501 runs the computer program stored in the memory 502, it can execute the above-mentioned method for determining the liveness of an individual.

[0133] Corresponding to the above-described method for identifying a living body, this application also provides a computer-readable storage medium storing a computer program, which, when run by a processor, executes the steps of the above-described method for identifying a living body.

[0134] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some communication interfaces; the indirect coupling or communication connection of devices or modules can be electrical, mechanical, or other forms.

[0135] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0136] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0137] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the information processing methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0138] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for identifying living organisms, characterized in that, The method for identifying living organisms includes: Acquire a video of a user reading and judging text; the lip shape category corresponding to the judging text is the same in different dialects; Obtain the mouth shape coordinate information corresponding to each video frame in the video of the determined text; All mouth shape coordinate information is sequentially input into the mouth shape detection model to obtain the mouth shape category of the user in the video; the mouth shape detection model is trained using mouth shape coordinate information samples and corresponding mouth shape categories. Determine whether the user's lip shape category is the same as the lip shape category corresponding to the judgment text; If the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text, then the user is determined to be alive.

2. The method for identifying living organisms according to claim 1, characterized in that, The mouth shape categories include the category of opening the mouth first and then closing it, the category of closing the mouth first and then opening it, and the category of pouting. The method further includes: Select one character from each of the preset characters corresponding to the categories of "open mouth then close mouth", "close mouth then open mouth", and "pouting mouth" and combine them as the judgment character for the user.

3. The method for identifying living organisms according to claim 1 or 2, characterized in that, The step of obtaining the lip shape coordinate information corresponding to each video frame in the video of the determined text includes: The initial lip shape coordinates of each video frame in the video are detected using a facial key point detection model. The initial mouth shape coordinate information is filtered using the Kalman filter algorithm to obtain the target mouth shape coordinate information.

4. The method for identifying living organisms according to claim 3, characterized in that, Before determining whether the user's lip shape category is the same as the lip shape category corresponding to the judgment text, the method further includes: Obtain the audio of the user reading the text to be judged; The audio is converted into text to obtain the reading text that the user is reading in the audio. After determining that the user's lip shape category is the same as the lip shape category corresponding to the judgment text, the method further includes: Determine whether the text being read by the user is the same as the text being judged; If the text being read by the user is the same as the text being judged, then the user is determined to be a living person.

5. The method for identifying living organisms according to claim 4, characterized in that, The step of converting the audio into text to obtain the reading text that the user is reading in the audio includes: The audio is converted into text using automatic speech recognition technology, resulting in the first reading text that the user is reading in the audio.

6. The method for identifying living organisms according to claim 5, characterized in that, The step of converting the audio into text to obtain the reading text of the user in the audio also includes: After adding Gaussian noise to the audio, the audio with added Gaussian noise is converted into text using automatic speech recognition technology to obtain the second reading text read by the user in the audio.

7. The method for identifying living organisms according to claim 6, characterized in that, The step of determining whether the text being read by the user is the same as the text being judged includes: If the first or second reading text is the same as the determined text, then the reading text read by the user is the same as the determined text.

8. A device for identifying living organisms, characterized in that, The liveness detection device includes: The acquisition module is used to acquire videos of users reading and judging text; the lip shape categories corresponding to the judging text are the same in different dialects; The acquisition module is also used to acquire the mouth shape coordinate information corresponding to each video frame in the video of the determined text; The input module is used to sequentially input all mouth shape coordinate information into the mouth shape detection model to obtain the mouth shape category of the user in the video; the mouth shape detection model is trained by mouth shape coordinate information samples and corresponding mouth shape categories. The judgment module is used to determine whether the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text; if the user's mouth shape category is the same as the mouth shape category corresponding to the judgment text, then the user is determined to be alive.

9. An electronic device, characterized in that, include: The device includes a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is in operation, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the liveness detection method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the method for determining a living body as described in any one of claims 1 to 7.

Images