Sample processing method, object detection method, identity verification method and related products

Abstract

The embodiment of the invention discloses a sample processing method, an object detection method, an identity verification method and a related product. The sample processing method comprises: according to a first time-frequency diagram of an ultrasonic signal in a sample, performing signal quality detection on the ultrasonic signal to obtain a signal quality detection result, the first time-frequency diagram being used for representing a frequency shift amplitude of the ultrasonic signal within a preset duration; under the condition that the signal quality detection result is that the preset condition is met, the maximum frequency shift amplitude of the ultrasonic signal in the preset duration is determined according to the first time-frequency diagram; and when the maximum frequency shift amplitude is greater than or equal to a preset frequency shift threshold, determining that the sample meets a sample screening condition. According to the method and the device, the optimization of the sample can be ensured, so that the model performance of the trained model can be improved when the sample is applied to model training.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a sample processing method, an object detection method, an identity authentication method and related products. Background Art

[0002] Detection systems based on deep neural network technology have been widely used in various front-end applications. For example, for financial business-related systems, detection systems are usually used in applications such as banking and payment. Before processing related business, the system is first used to detect the authenticity of the identity, and after verification, the business process is further executed to avoid the application from being attacked. It can be seen that the performance of the detection system is closely related to the processing results of the front-end business. How to improve the detection performance of the system is a crucial issue. Summary of the invention

[0003] The purpose of the embodiments of the present application is to provide a sample processing method, an object detection method, an identity authentication method and related products to ensure the quality of the samples, so that when the samples are applied to the training of the first model, the model performance of the trained first model can be improved.

[0004] To solve the above technical problems, the embodiments of the present application are implemented as follows: On the one hand, an embodiment of the present application provides a sample processing method, comprising: According to a first time-frequency graph of an ultrasonic signal in a sample, a signal quality detection is performed on the ultrasonic signal to obtain a signal quality detection result, wherein the ultrasonic signal is an ultrasonic signal extracted from the sample, and the first time-frequency graph is used to represent a frequency shift amplitude of the ultrasonic signal within a preset time length; When the signal quality detection result satisfies a preset condition, determining a maximum frequency shift amplitude of the ultrasonic signal within the preset time length according to the first time-frequency graph; When the maximum frequency shift amplitude is greater than or equal to a preset frequency shift threshold, it is determined that the sample meets the sample screening condition.

[0005] On the other hand, an embodiment of the present application provides an object detection method, including: According to a time-frequency diagram of a first ultrasonic signal of the object, a signal quality detection is performed on the first ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the first ultrasonic signal within a preset time length; In the case where the signal quality detection result satisfies the preset condition, determining the maximum frequency shift amplitude of the first ultrasonic signal within the preset time length according to the time-frequency diagram; The object is detected according to the first ultrasonic signal and the maximum frequency shift amplitude to obtain a detection result of the object.

[0006] On the other hand, the embodiment of the present application provides an identity authentication method, including: According to the time-frequency diagram of the second ultrasonic signal of the human face, a signal quality detection is performed on the second ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the second ultrasonic signal within a preset time length; In the case where the signal quality detection result satisfies the preset condition, determining the maximum frequency shift amplitude of the second ultrasonic signal within the preset time length according to the time-frequency diagram; The human face is authenticated according to the second ultrasonic signal and the maximum frequency shift amplitude to obtain an identity authentication result of the human face.

[0007] On the other hand, an embodiment of the present application provides a sample processing device, including: a first detection module, configured to perform a signal quality detection on the ultrasonic signal according to a first time-frequency graph of the ultrasonic signal in the sample to obtain a signal quality detection result, wherein the ultrasonic signal is an ultrasonic signal extracted from the sample, and the first time-frequency graph is used to represent a frequency shift amplitude of the ultrasonic signal within a preset time length; A first determination module, configured to determine, when the signal quality detection result satisfies a preset condition, a maximum frequency shift amplitude of the ultrasonic signal within the preset time length according to the first time-frequency graph; The second determination module is used to determine that the sample meets the sample screening condition when the maximum frequency shift amplitude is greater than or equal to a preset frequency shift threshold.

[0008] On the other hand, an embodiment of the present application provides an object detection device, including: A second detection module is used to perform a signal quality detection on the first ultrasonic signal of the object according to a time-frequency diagram of the first ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the first ultrasonic signal within a preset time length; A third determination module is used to determine, when the signal quality detection result satisfies a preset condition, a maximum frequency shift amplitude of the first ultrasonic signal within the preset time length according to the time-frequency diagram; The third detection module is used to detect the object according to the first ultrasonic signal and the maximum frequency shift amplitude to obtain a detection result of the object.

[0009] On the other hand, an embodiment of the present application provides an identity authentication device, including: a fourth detection module, configured to perform a signal quality detection on the second ultrasonic signal of the face according to a time-frequency diagram of the second ultrasonic signal, and obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the second ultrasonic signal within a preset time length; a fourth determination module, configured to determine, if the signal quality detection result satisfies a preset condition, a maximum frequency shift amplitude of the second ultrasonic signal within the preset time length according to the time-frequency diagram; An identity authentication module is used to authenticate the face according to the second ultrasonic signal and the maximum frequency shift amplitude to obtain an identity authentication result of the face.

[0010] On the other hand, an embodiment of the present application provides an electronic device, comprising a processor and a memory electrically connected to the processor, the memory storing a computer program, the processor being used to call and execute the computer program from the memory to implement the above-mentioned sample processing method, or the processor being used to call and execute the computer program from the memory to implement the above-mentioned object detection method, or the processor being used to call and execute the computer program from the memory to implement the above-mentioned identity authentication method.

[0011] On the other hand, an embodiment of the present application provides a computer-readable storage medium for storing a computer program, wherein the computer program can be executed by a processor to implement the above-mentioned sample processing method, or the computer program can be executed by a processor to implement the above-mentioned object detection method, or the computer program can be executed by a processor to implement the above-mentioned identity authentication method.

[0012] On the other hand, an embodiment of the present application provides a computer program product, including a computer program, wherein the computer program is executed by a processor to implement the above-mentioned sample processing method, or the computer program is executed by a processor to implement the above-mentioned object detection method, or the computer program is executed by a processor to implement the above-mentioned identity authentication method.

[0013] According to the technical solution of the embodiment of the present application, the signal quality detection of the ultrasonic signal is performed according to the first time-frequency diagram of the ultrasonic signal in the sample to obtain the signal quality detection result; when the signal quality detection result satisfies the preset condition, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the first time-frequency diagram; when the maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, it is determined that the sample meets the sample screening condition. It can be seen that the acquired sample is not used directly, for example, it is not directly applied to the model training scene, but the signal quality detection of the ultrasonic signal in the sample is performed in advance to ensure that the signal quality of the ultrasonic signal is qualified. In addition, when the signal quality of the ultrasonic signal is qualified, the frequency shift amplitude of the ultrasonic signal is further detected. Since a large image area shows different modes when the frequency shift is large enough, it is beneficial for the later model to learn the difference between positive and negative samples. If the frequency shift is very small, it can be considered that the target object corresponding to the sample has no movement, which does not conform to the principle of the frequency shift phenomenon caused by the movement of the target object in this solution. Therefore, by ensuring that the processed ultrasonic signal has sufficient frequency shift, when using the sample, for example, when the sample is applied to the training first model scene, the sample is ensured to be high-quality, and the model performance of the trained first model is further improved, thereby improving the detection performance of the detection system. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate one or more embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in one or more embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0015] Figure 1 is a schematic flow chart of a sample processing method according to an embodiment of the present application; Figure 2 is a schematic interface diagram of a time-frequency diagram of an ultrasound signal according to an embodiment of the present application; Figure 3 is a schematic interface diagram of a time-frequency diagram of an ultrasound signal according to another embodiment of the present application; Figure 4 is a schematic interface diagram of a time-frequency diagram of an ultrasonic signal according to yet another embodiment of the present application; Figure 5 is a schematic interface diagram of a time-frequency diagram of an ultrasonic signal according to yet another embodiment of the present application; Figure 6 is a schematic flow chart of a sample processing method according to another embodiment of the present application; Figure 7is a schematic flow chart of an object detection method according to an embodiment of the present application; Figure 8 is a schematic flow chart of an identity authentication method according to an embodiment of the present application; Fig. 9 is a schematic block diagram of a sample processing device according to an embodiment of the present application; Fig.10 is a schematic block diagram of an object detection device according to an embodiment of the present application; Fig.11 is a schematic block diagram of an identity authentication device according to an embodiment of the present application; Fig.12 It is a schematic block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0016] The embodiments of the present application provide a sample processing method, an object detection method, an identity authentication method and related products to ensure the quality of the samples, so that when the samples are applied to the training of the first model, the model performance of the trained first model can be improved.

[0017] In order to enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

[0018] In terms of object detection, in the related art, forgery detection is usually performed based on the pixel texture analysis technology of two-dimensional images. For example, images with high pixels are considered to be provided by real users, while images with low pixels are considered to be forged by attackers. Although this pixel texture analysis technology has made great progress, it is limited by ambient light, pixel texture and other limitations, and it is difficult to adapt to increasingly complex attack methods. To this end, the related art proposes a technology that combines ultrasonic technology and deep neural networks for object detection. Specifically, an ultrasonic signal is emitted through a speaker in an electronic device, and a reflected echo signal is received through a microphone, and the reflected echo signal is used as sample data to train an object detection model. The trained model can be applied to object detection verification scenarios, by collecting video when the user inputs action information, extracting ultrasonic signals from the video, and then analyzing the collected ultrasonic signals to determine whether the current verification scenario is an attack.

[0019] However, in actual application scenarios, verification based on ultrasonic signals (such as object detection) still faces great challenges. On the one hand, the speakers and microphones in electronic devices (such as mobile phones) do not have strict quality control in the ultrasonic band (between 16kHz and 22kHz), resulting in unstable operation of the speakers and microphones. This instability varies depending on the quality of the electronic equipment. For example, sometimes it manifests as insufficient energy in the transmission frequency of the ultrasonic signal, and sometimes it manifests as insufficient sampling frequency shift amplitude of the ultrasonic signal, etc. On the other hand, it is difficult to strictly regulate the user's input of action information (such as nodding, shaking head, etc.). For example, the user's action amplitude is not enough, the mobile phone is not facing the face, the distance is too far, etc., which will cause the ultrasonic signal collected by the mobile phone to be extremely weak and the frequency shift is insufficient, resulting in the unqualified quality of the collected ultrasonic signal, which cannot be used for model training. If this type of ultrasonic signal is used for model training, the performance of the trained model will be poor, and the verification effect of the model's detection effect for object detection cannot be ensured.

[0020] In the sample processing method provided by the present application, the signal quality detection of the ultrasonic signal is performed according to the first time-frequency diagram of the ultrasonic signal in the sample to obtain the signal quality detection result; when the signal quality detection result satisfies the preset condition, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the first time-frequency diagram; when the maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, it is determined that the sample meets the sample screening condition. It can be seen that the acquired sample is not used directly, for example, it is not directly applied to the model training scene, but the signal quality detection of the ultrasonic signal in the sample is performed in advance to ensure that the signal quality of the ultrasonic signal is qualified. In addition, when the signal quality of the ultrasonic signal is qualified, the frequency shift amplitude of the ultrasonic signal is further detected. Since a larger image area shows different modes when the frequency shift is large enough, it is beneficial for the later model to learn the difference between positive and negative samples. If the frequency shift is very small, it can be considered that the target object corresponding to the sample has no movement, which does not conform to the principle of the frequency shift phenomenon caused by the movement of the target object in this scheme. Therefore, by ensuring that the processed ultrasonic signal has sufficient frequency shift, when the sample is used, for example, when the sample is applied to the training liveness detection model scene, the quality of the sample is ensured, and the model performance of the trained first model is further improved.

[0021] The sample processing method, object detection method and identity authentication method provided in the embodiments of the present application can be executed by an electronic device, or by software installed in the electronic device. Specifically, the electronic device can be a terminal device or a server device. Among them, the terminal device can include a smart phone, a laptop computer, a smart wearable device, a vehicle terminal, etc., and the server device can include an independent physical server, a server cluster composed of multiple servers, or a cloud server capable of cloud computing.

[0022] Figure 1 is a schematic flow chart of a sample processing method according to an embodiment of the present application, such as Figure 1 As shown, the sample processing method includes the following steps: S102, performing a signal quality detection on the ultrasonic signal according to a first time-frequency diagram of the ultrasonic signal in the sample to obtain a signal quality detection result; the ultrasonic signal is an ultrasonic signal extracted from the sample, and the first time-frequency diagram is used to represent the frequency shift amplitude of the ultrasonic signal within a preset time length.

[0023] Optionally, the sample is represented by a video, and the ultrasonic signal can be extracted from the video. The video (i.e., the sample) can be acquired through an electronic device. For example, the facial movements of the user can be acquired through the camera function of the electronic device to obtain a video generated by the facial movements. The video (i.e., the sample) can also be acquired through virtual software installed on the electronic device. For example, the facial movements of the user can be first acquired through the video generation function of the virtual software, and then the corresponding video can be generated according to the facial movement information. In the case where the sample includes multiple videos, the ultrasonic signal is extracted from each video respectively, so that each video corresponds to its own ultrasonic signal.

[0024] Optionally, the ultrasonic signal in the sample can be obtained through the speaker and microphone in the electronic device. Taking the electronic device as a terminal device as an example, in the process of object detection, for example, for object detection for gesture recognition, similarly, the terminal device is aimed at the gesture action process, and the gesture action information (such as waving hands left and right or waving hands forward and backward) is input. At this time, the speaker in the terminal device transmits an ultrasonic signal, and the microphone receives the echo signal of the ultrasonic signal. The ultrasonic signal in the sample is the echo signal received by the microphone. After the terminal device transmits an ultrasonic signal to the gesture action, the gesture action will cause the ultrasonic signal to shift in frequency. The larger the amplitude of the gesture action, the larger the frequency shift amplitude of the ultrasonic signal. Therefore, in order to make the collected ultrasonic signal have a larger frequency shift amplitude, the gesture action can select an action type with a larger amplitude, while for a gesture action with a smaller amplitude, it may cause the collected ultrasonic signal to have insufficient frequency shift. In addition, in the process of object detection, the activity of animals and the liveness recognition of animals can also be detected.

[0025] Taking the electronic device as a mobile phone as an example, when performing face identity authentication, the user holds the mobile phone so that the mobile phone is facing the face and enters the facial action information (such as shaking the head or nodding). At this time, the speaker in the mobile phone transmits an ultrasonic signal, and the microphone receives the echo signal of the ultrasonic signal. The sample ultrasonic signal is the echo signal received by the microphone. Optionally, the speaker can transmit the ultrasonic signal at a fixed frequency. Usually, the frequency of the ultrasonic signal is between 16KHz and 22KHz. After the electronic device transmits the ultrasonic signal to the face, the facial action will cause the ultrasonic signal to shift in frequency. The larger the amplitude of the facial action, the larger the frequency shift amplitude of the ultrasonic signal. Therefore, in order to make the collected ultrasonic signal have a larger frequency shift amplitude, the facial action can select an action type with a larger amplitude, such as nodding or shaking the head, while for facial actions with a smaller amplitude (such as blinking, opening the mouth, etc.), the collected ultrasonic signal may have insufficient frequency shift.

[0026] Taking the face identity authentication scenario as an example, the preset duration can be determined according to the duration of the face identity authentication. Optionally, the preset duration is less than or equal to the duration of the user inputting the facial motion information. After triggering the electronic device to perform face identity authentication, the electronic device prompts the user to input the facial motion information, and starts collecting ultrasonic signals at the same time. After obtaining the facial motion information input by the user, the ultrasonic signal collection stops. Alternatively, the preset duration for collecting ultrasonic signals can be pre-set. For example, if the preset duration is 2 seconds, the electronic device collects ultrasonic signals within 2 seconds after the electronic device prompts the user to input the facial motion information.

[0027] The first time-frequency diagram of the ultrasonic signal can be calculated based on the short-time Fourier transform, and when the first time-frequency diagram is calculated, the spectrum energy value of each pixel point in the first time-frequency diagram can be calculated based on the short-time Fourier transform. Since the process of calculating the time-frequency diagram and the spectrum energy value of each pixel point in the time-frequency diagram based on the short-time Fourier transform is a mature technology, it will not be repeated here.

[0028] Optionally, when performing signal quality detection on an ultrasonic signal, detection may be performed from at least one of the following aspects: whether the spectrum energy of the central region of the ultrasonic signal is large enough, and whether the background noise of the ultrasonic signal is small enough. Only when the spectrum energy of the central region of the ultrasonic signal is small, or the background noise of the ultrasonic signal is large, will it be determined that the signal quality detection of the ultrasonic signal has passed. The central region is the region where the center frequency is located.

[0029] If the spectral energy in the center area of ​​the ultrasonic signal is small, it means that the energy of the central transmission frequency of the ultrasonic signal is insufficient. The reason may be that there is a quality problem with the electronic device or the electronic device is far away from the face, which will lead to low discriminability of the ultrasonic signal. The sample corresponding to the ultrasonic signal cannot be used as sample data for model training. Low discriminability means that the accuracy of ultrasonic signals being judged as positive sample signals or negative sample signals is low.

[0030] If the background noise of the ultrasonic signal is large, it means that the electronic device collecting the ultrasonic signal may be in an unstable state. For example, the electronic device is frequently shaken, which will cause the noise in the ultrasonic signal to be too large and the signal type cannot be effectively identified. Therefore, it cannot be used as sample data for model training.

[0031] S104: When the signal quality detection result satisfies a preset condition, a maximum frequency shift amplitude of the ultrasonic signal within a preset time length is determined according to the first time-frequency graph.

[0032] The first time-frequency diagram is a diagram of the relationship between time and frequency. Optionally, the first time-frequency diagram is represented in the form of a coordinate diagram. Assuming that an ultrasonic signal is emitted at a center frequency of 17200 Hz, the time-frequency diagram of the collected ultrasonic signal can be represented as Figure 2 As shown, the horizontal axis of the coordinate graph is the time dimension, the unit is seconds (s), and the vertical axis is the frequency dimension, the unit is Hz. Each row in the first video graph corresponds to the spectrum energy value of a frequency, as shown on the right side of the coordinate graph. Figure 2 It can be seen that each row corresponds to the spectrum energy value of a frequency, and near the row where the center frequency f=17200Hz is located, there is a higher spectrum energy value. Figure 2 It appears as a brighter area in the middle.

[0033] S106: When the maximum frequency shift amplitude is greater than or equal to a preset frequency shift threshold, determine that the sample meets the sample screening condition.

[0034] Among them, the sample meets the sample screening conditions, which means that the signal quality of the ultrasonic signal in the sample is high and has sufficient frequency shift amplitude, and can be used as sample data for training the first model.

[0035] The maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, indicating that the ultrasonic signal has sufficient frequency shift and is therefore highly discriminable. The sample category of the sample may include a positive sample category or a negative sample category. The ultrasonic signal in the sample of the positive sample category is collected when the real user provides facial motion information. The ultrasonic signal in the sample of the negative sample category is an attack signal. For example, when an attacker performs facial identity authentication through photos, paper, or electronic screen playback, the ultrasonic signal collected from the attacker is an attack signal.

[0036] The sample is processed based on the sample processing method provided in this embodiment to obtain a processed sample. The processed sample can be applied to a variety of scenarios for detection based on object motion. The object can be any movable object such as a person, an animal, or a vehicle. For example, when the object is a person, the sample can be facial motion data (such as a video), and the processed sample can be applied to a scenario for identity authentication based on facial motion (such as shaking the head, nodding, etc.). When the object is an animal, the sample can be animal body motion data (such as a video), and the processed sample can be applied to a scenario for animal species identification based on animal body motion (such as running, walking, jumping, etc.). When the object is a vehicle, the sample can be vehicle motion data (such as a video), and the processed sample can be applied to a scenario for vehicle performance detection based on vehicle motion (such as turning around, steering, going straight, etc.).

[0037] The processed samples can also be applied to model training in a variety of scenarios, such as object detection scenarios (such as identification of animal types, detection of vehicle performance, etc.), identity verification scenarios, etc. For samples of different scenarios, after sample processing, the obtained samples can be used to train different types of models. Exemplarily, in an object detection scenario, the samples can be applied to training an object detection model after processing, and the trained object detection model can be used to perform object detection. For example, when the object is an animal, the processed samples are used to train an animal species recognition model; when the object is a vehicle, the processed samples are used to train a vehicle performance detection model. In an identity verification scenario, the samples can be applied to training a face recognition model after processing, and the trained face recognition model can be applied to user identity verification.

[0038] Of course, the above content is only an exemplary description of the scenario to which the sample belongs, and does not constitute a limitation on the scenario to which the sample belongs. Accordingly, the type of model is also not limited.

[0039] According to the technical solution of the embodiment of the present application, the signal quality detection of the ultrasonic signal is performed according to the first time-frequency diagram of the ultrasonic signal in the sample to obtain the signal quality detection result; when the signal quality detection result satisfies the preset condition, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the first time-frequency diagram; when the maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, it is determined that the sample meets the sample screening condition. It can be seen that the acquired sample is not used directly, for example, it is not directly applied to the model training scene, but the signal quality detection of the ultrasonic signal in the sample is performed in advance to ensure that the signal quality of the ultrasonic signal is qualified. In addition, when the signal quality of the ultrasonic signal is qualified, the frequency shift amplitude of the ultrasonic signal is further detected. Since a large image area shows different modes when the frequency shift is large enough, it is beneficial for the later model to learn the difference between positive and negative samples. If the frequency shift is very small, it can be considered that the target object corresponding to the sample has no movement, which does not conform to the principle of the frequency shift phenomenon caused by the movement of the target object in this solution. Therefore, by ensuring that the processed ultrasonic signal has sufficient frequency shift, when the sample is used, for example, when the sample is applied to the training liveness detection model scene, the quality of the sample is ensured, and the model performance of the trained first model is further improved.

[0040] In one embodiment, when the ultrasonic signal satisfies at least one of the following conditions A1-A2, it is determined that the signal quality detection result of the ultrasonic signal satisfies the preset condition: Condition A1: an average value of spectral energy of the ultrasonic signal in the first center offset region of the first time-frequency diagram is greater than or equal to a preset energy threshold.

[0041] The offset of the pixel points in the first central offset region relative to the central region of the first time-frequency graph is less than or equal to a preset offset threshold. The central region is the region where the central frequency is located. Optionally, the central region is the row where the central frequency is located. Figure 2 Taking the time-frequency diagram shown as an example, the center frequency f=17200 Hz, and the central area is the row where the center frequency f is located.

[0042] The first center offset region is a region including the center region and whose distance from the center region does not exceed a preset offset threshold, expressed as [fC, f+C], where f is the center frequency and C is the preset offset threshold. Figure 2 Taking the time-frequency diagram shown as an example, the center frequency f=17200 Hz, and the center area is the row where the center frequency f is located. Assuming that the preset offset threshold is 8 Hz, the first center offset area is the range of 8 Hz above and below the center frequency f, expressed as [f-8, f+8].

[0043] The average value of the spectrum energy of the first center offset region refers to the average value of the spectrum energy values ​​of each pixel point in the first center offset region.

[0044] Condition A2: The background noise of the ultrasonic signal meets the preset noise condition.

[0045] Optionally, when determining whether the background noise of the ultrasonic signal meets the preset noise condition, the first time-frequency graph may be firstly binarized according to the spectrum energy value of each pixel point in the first time-frequency graph to obtain a second time-frequency graph, and then the first connected domain in the second time-frequency graph is determined. Thereafter, when the number of the first connected domains is less than or equal to the preset connected domain threshold, it is determined that the background noise meets the preset noise condition.

[0046] Among them, when the first time-frequency graph is binarized, the spectral energy value of the pixel point can be compared with the preset energy threshold. If the spectral energy value of the pixel point is greater than the preset energy threshold, the value of the pixel point is set to 1. If the spectral energy value of the pixel point is less than or equal to the preset energy threshold, the value of the pixel point is set to 0. By setting the value of each pixel point in the first time-frequency graph to 0 or 1, the first video image is processed into a binary image including only 0 and 1, that is, the second time-frequency graph. In the second time-frequency graph, a pixel point with a value of 0 indicates that the spectral energy value of the pixel point is small, which is displayed as black. A pixel point with a value of 1 indicates that the spectral energy value of the pixel point is large, which is displayed as white.

[0047] Figure 3 is a schematic interface diagram of a second time-frequency diagram according to an embodiment of the present application, such as Figure 3 As shown in the figure, the first connected domain is the white connected domain in the second time-frequency diagram. If the number of first connected domains is large, for example, more than the preset connected domain threshold, it means that the background noise of the ultrasound signal is too large and the discriminability is low, and the sample corresponding to the ultrasound signal is not suitable for training the model. For the convenience of display, Figure 3 The first connected domain is represented by a circular area. In practical applications, the first connected domain is usually irregular in shape. The shaded area represents an area with low spectrum energy value. In practical applications, the shaded area appears as a dark area.

[0048] In this embodiment, only when the average value of the spectrum energy of the first center offset area of ​​the ultrasonic signal in the first time-frequency diagram is greater than or equal to the preset energy threshold, and / or the background noise of the ultrasonic signal meets the preset noise condition, will the signal quality detection result of the ultrasonic signal be determined to meet the preset condition, thereby ensuring that the signal quality of the ultrasonic signal is qualified. If the samples corresponding to the ultrasonic signal whose signal quality detection result meets the preset condition are used to train the first model, that is, by pre-screening the samples corresponding to the ultrasonic signal with qualified signal quality as the training samples of the first model, the sample quality of the samples used to train the first model can be made better, thereby improving the model performance of the trained first model.

[0049] In one embodiment, when determining the maximum frequency shift amplitude of the ultrasonic signal within a preset time length according to the first time-frequency graph, it can be determined by a binarized time-frequency graph. Optionally, the maximum frequency shift amplitude is still determined by using the second time-frequency graph obtained after binarization in the above embodiment. First, the maximum connected domain in the second time-frequency graph is determined, and then the maximum offset of the maximum connected domain relative to the central area of ​​the second time-frequency graph is determined, and the maximum offset is determined as the maximum frequency shift amplitude of the ultrasonic signal within the preset time length.

[0050] In the second time-frequency graph, a pixel with a value of 0 indicates that the spectral energy value of the pixel is small and is displayed in black. A pixel with a value of 1 indicates that the spectral energy value of the pixel is large and is displayed in white. Therefore, the maximum connected domain in the second time-frequency graph is the white connected domain with the largest area.

[0051] The maximum offset of the maximum connected domain relative to the central area of ​​the second time-frequency graph can be represented as the distance between the pixel point farthest from the central area in the maximum connected domain and the central area. Since the central area is the row where the central frequency is located, the distance between the pixel point farthest from the central area in the maximum connected domain and the central area is the distance between the pixel point farthest from the central area in the maximum connected domain and the row where the central frequency is located.

[0052] like Figure 4 As shown, the maximum connected domain is region S, and the central region is the row where the central frequency is located, denoted as X. When calculating the maximum offset of the maximum connected domain S relative to the central region X, the pixel point farthest from the central region X in the maximum connected domain S can be first determined, for example Figure 4 As shown in the figure, the distance P between the pixel point A and the central area X is the maximum offset of the maximum connected domain S relative to the central area X, that is, the maximum frequency shift amplitude of the sample ultrasonic signal within the preset time length.

[0053] In one embodiment, the sample carries label information, which can be any form of information that can characterize the positive sample category or the negative sample category. For example, the label information is represented in digital form, 0 represents the negative sample category, and 1 represents the positive sample category; or the label information is represented in the form of keywords, "positive" represents the negative sample category, and "negative" represents the positive sample category; and so on.

[0054] After determining that the sample meets the sample screening condition, the sample can be used to train the first model. Optionally, the sample category of the sample is determined based on the label information carried by the sample, and the sample category includes a positive sample category and a negative sample category; and then the first model is trained based on the sample and the corresponding sample category.

[0055] In one embodiment, the sample is discarded if the ultrasonic signal in the sample meets at least one of the following conditions: the average spectral energy value of the ultrasonic signal in the first center offset region is less than a preset energy threshold, the background noise of the ultrasonic signal does not meet the preset noise condition, and the maximum frequency shift amplitude of the ultrasonic signal within a preset time length is less than a preset frequency shift threshold.

[0056] By discarding samples in any of the above situations, samples corresponding to ultrasonic signals with unqualified signal quality and insufficient frequency shift can be screened out, thereby ensuring that the ultrasonic signals in the processed samples not only have qualified signal quality but also have sufficient frequency shift and high discriminability. If the samples obtained after screening are used to train the first model, the model performance of the trained first model can be further improved.

[0057] In one embodiment, before performing signal quality detection on the ultrasound signal according to the first time-frequency graph, the second center offset region in the first video image may be shielded to obtain a shielded time-frequency graph. Then, the signal quality detection is performed on the ultrasound signal according to the shielded time-frequency graph.

[0058] When obtaining the time-frequency diagram after shielding, the second center offset region of the ultrasonic signal in the first time-frequency diagram is first determined, and then the second center offset region is shielded to obtain the time-frequency diagram after shielding. The second center offset region includes the first center offset region. The second center offset region can be the first center offset region, or a region larger than the first center offset region. After the second center offset region is shielded, the region appears black, i.e., the darkest region, such as Figure 5 The area shown is located near the center frequency of 17200 Hz.

[0059] Still Figure 2 Taking the first time-frequency diagram shown as an example, assuming that the first center offset region in the first time-frequency diagram is represented as [fC, f+C], f is the center frequency, and C is the preset offset threshold. Then, the second center offset region can be represented as [fC, f+C], that is, the second center offset region is the same as the first center offset region; or, the second center offset region can be represented as [f-C1, f+C1], C1 is greater than C, that is, the second center offset region is a region greater than the first center offset region. For example, if C is 8Hz, C1 can be determined as other values ​​greater than 8Hz, such as 10Hz or 15Hz, etc.

[0060] In addition, when shielding the second center offset area, the entire area of ​​the second center offset area can be shielded, or each row can be shielded or not shielded according to the spectrum energy value of each row in the second center offset area. The spectrum energy value of each row is the average of the spectrum energy values ​​of each pixel point in the row. For any row in the second center offset area, if the spectrum energy value of the row is greater than or equal to the preset energy threshold, the row is shielded; if the spectrum energy value of the row is less than the preset energy threshold, the row is not shielded.

[0061] In this embodiment, the second center offset area is the main transmission frequency area of ​​the ultrasonic signal. The spectral energy value of the pixel points in this area is very high, which may be higher than the frequency offset (i.e., frequency shift) caused by the movement of the target object. Therefore, in order to accurately determine the frequency offset caused by the movement of the target object, this embodiment shields the second center offset area to avoid the frequency offset caused by the movement of the target object being covered by the second center offset area.

[0062] Figure 6 is a schematic flow chart of a sample processing method according to another embodiment of the present application. In this embodiment, the processed samples are used to train a liveness detection model, and the liveness detection model is applied in a face identity verification scenario. Figure 6 As shown, the sample processing method includes the following steps S601-S608: S601, acquiring an ultrasonic signal in a sample and a first time-frequency diagram of the ultrasonic signal.

[0063] Among them, the ultrasonic signal is obtained through the speaker and microphone in the electronic device. Taking the electronic device as a mobile phone as an example, when performing face identity authentication, the user holds the mobile phone so that the mobile phone is facing the face and enters the facial movement information (such as shaking the head or nodding). At this time, the speaker in the mobile phone emits an ultrasonic signal at a fixed frequency, and the microphone receives the echo signal of the ultrasonic signal. The sample ultrasonic signal is the echo signal received by the microphone, and the frequency of the ultrasonic signal emitted by the speaker is between 16KHz and 22KHz.

[0064] The first time-frequency diagram is used to characterize the frequency shift amplitude of the ultrasonic signal within a preset time length. In the first time-frequency diagram, the horizontal axis is the time dimension, the vertical axis is the frequency dimension, and each row corresponds to the spectrum energy value of a frequency. The coordinate diagram representation of the first time-frequency diagram can be referred to Figure 2 As shown, it will not be repeated here.

[0065] S602: Determine an average value of spectrum energy of a first center offset region of the ultrasonic signal in a first time-frequency diagram.

[0066] The offset of the pixel points in the first central offset region relative to the central region of the first time-frequency graph is less than or equal to the preset offset threshold. The central region is the row where the central frequency is located. Figure 2 Taking the time-frequency diagram shown as an example, the center frequency f=17200Hz, and the center area is the row where the center frequency f is located. The first center offset area can be expressed as [fC, f+C], where C is the preset offset threshold. Assuming the preset offset threshold is 8Hz, the first center offset area is the range of 8Hz above and below the center frequency f, expressed as [f-8, f+8].

[0067] S603, when the average spectral energy of the ultrasonic signal in the first central offset region is greater than or equal to the first preset energy threshold, determine the second central offset region of the ultrasonic signal in the first time-frequency diagram, and shield the high energy region in the second central offset region to obtain the shielded first time-frequency diagram.

[0068] Among them, the high energy area is in units of rows, and the high energy area refers to the rows in the second center offset area whose spectrum energy value is greater than or equal to the second preset energy threshold. Specifically, each row in the second center offset area is shielded or not shielded respectively. The spectrum energy value of each row is the average of the spectrum energy values ​​of each pixel point in the row. For any row in the second center offset area, if the spectrum energy value of the row is greater than or equal to the second preset energy threshold, the row is shielded; if the spectrum energy value of the row is less than the second preset energy threshold, the row is not shielded.

[0069] The second center offset area includes the first center offset area. The second center offset area can be the first center offset area, or it can be an area larger than the first center offset area. After the second center offset area is shielded, the area appears black, that is, the darkest area. Figure 5 shown.

[0070] When the average energy value of the spectrum of the ultrasonic signal in the first center offset region is less than the first preset energy threshold, it indicates that the energy of the center transmission frequency of the ultrasonic signal is insufficient and the sample can be discarded.

[0071] S604: binarize the masked first time-frequency graph according to the spectrum energy value of each pixel point in the masked first time-frequency graph to obtain a second time-frequency graph.

[0072] Among them, when the first time-frequency graph is binarized to obtain the second time-frequency graph, the spectral energy value of the pixel point can be compared with the third preset energy threshold. If the spectral energy value of the pixel point is greater than the third preset energy threshold, the value of the pixel point is set to 1. If the spectral energy value of the pixel point is less than or equal to the third preset energy threshold, the value of the pixel point is set to 0. By setting the value of each pixel point in the first time-frequency graph to 0 or 1, the first video image is processed into a binary image including only 0 and 1, that is, the second time-frequency graph. In the second time-frequency graph, a pixel point with a value of 0 indicates that the spectral energy value of the pixel point is small, which is displayed as black. A pixel point with a value of 1 indicates that the spectral energy value of the pixel point is large, which is displayed as white.

[0073] S605: Determine the number of first connected regions in the second time-frequency graph.

[0074] S606: When the number of the first connected domains is less than or equal to the preset connected domain threshold, binarize the masked first time-frequency graph according to the spectrum energy value of each pixel point in the masked first time-frequency graph to obtain a third time-frequency graph.

[0075] Among them, when the first time-frequency graph is binarized to obtain the third time-frequency graph, the spectral energy value of the pixel point can be compared with the fourth preset energy threshold. If the spectral energy value of the pixel point is greater than the fourth preset energy threshold, the value of the pixel point is set to 1. If the spectral energy value of the pixel point is less than or equal to the fourth preset energy threshold, the value of the pixel point is set to 0. By setting the value of each pixel point in the first time-frequency graph to 0 or 1, the first video image is processed into a binary image including only 0 and 1, that is, the third time-frequency graph. In the third time-frequency graph, a pixel point with a value of 0 indicates that the spectral energy value of the pixel point is small, which is displayed as black. A pixel point with a value of 1 indicates that the spectral energy value of the pixel point is large, which is displayed as white.

[0076] Since in actual application scenarios, the signal frequency shift amplitude caused by the action of the target object is usually greater than the signal frequency shift caused by excessive background noise, the fourth preset energy threshold is usually determined to be a value greater than the third preset energy threshold. Of course, the fourth preset energy threshold may also be the same as the third preset energy threshold. In this case, there is no need to perform step S606, and the second time-frequency diagram in step S604 can be directly used as the third time-frequency diagram.

[0077] When the number of the first connected domains is greater than a preset connected domain threshold, the samples may be discarded.

[0078] S607, calculating the maximum offset of the maximum connected domain in the third time-frequency graph relative to the central area of ​​the third time-frequency graph, where the maximum offset is the maximum frequency shift amplitude of the ultrasonic signal within a preset time length.

[0079] S608: When the maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, the first model is trained according to samples corresponding to the ultrasonic signal.

[0080] Wherein, when the maximum frequency shift amplitude is smaller than a preset frequency shift threshold, the sample is discarded.

[0081] By adopting the technical solution of the embodiment of the present application, when training the first model, the ultrasonic signal in the acquired sample is not directly used for model training, but the signal quality detection of the ultrasonic signal in the sample is performed in advance, including the detection of the average value of the spectrum energy in the center offset area and the detection of the background noise of the ultrasonic signal, so as to ensure that the signal quality of the ultrasonic signal in the sample used for training the model is qualified. Moreover, when the signal quality of the ultrasonic signal is qualified, the frequency shift amplitude of the ultrasonic signal is further detected. Only when the maximum frequency shift amplitude of the ultrasonic signal is greater than or equal to the preset frequency shift threshold, the sample is used for model training to ensure that the ultrasonic signal in the sample used for training the model has sufficient frequency shift. In this way, the quality of the sample used for training the first model is ensured, and the model performance of the trained first model is further improved.

[0082] Figure 7 is a schematic flow chart of an object detection method according to an embodiment of the present application, such as Figure 7 As shown, the object detection method includes the following steps: S702: Perform a signal quality detection on the first ultrasonic signal according to a time-frequency diagram of the first ultrasonic signal of the object to obtain a signal quality detection result, wherein the time-frequency diagram is used to represent a frequency shift amplitude of the first ultrasonic signal within a preset time length. S704: When the signal quality detection result satisfies a preset condition, determine a maximum frequency shift amplitude of the first ultrasonic signal within a preset time length according to the time-frequency diagram.

[0083] S706: Detect the object according to the first ultrasonic signal and the maximum frequency shift amplitude to obtain a detection result of the object.

[0084] By adopting the technical solution of the embodiment of the present application, the signal quality detection of the ultrasonic signal is performed according to the time-frequency diagram of the first ultrasonic signal of the object to obtain the signal quality detection result; when the signal quality detection result meets the preset conditions, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the time-frequency diagram; and then the object is detected according to the ultrasonic signal and the maximum frequency shift amplitude to obtain the detection result of the object. It can be seen that when detecting the object, the acquired ultrasonic signal is not directly used for detection, but the signal quality detection of the ultrasonic signal is performed in advance to ensure that the signal quality of the acquired ultrasonic signal is qualified. Moreover, when the signal quality of the ultrasonic signal is qualified, it is further detected according to the frequency shift amplitude of the ultrasonic signal, which improves the quality of the ultrasonic signal used for detection and ensures the accuracy of detection based on the ultrasonic signal.

[0085] In one embodiment, when detecting an object according to the first ultrasonic signal and the maximum frequency shift amplitude and obtaining a detection result of the object, the following method may be performed: When the maximum frequency shift amplitude is less than the preset frequency shift threshold, the detection result is determined to be a detection failure. If the maximum frequency shift amplitude is less than the preset frequency shift threshold, it means that it does not cause sufficient frequency shift and can be considered as an attack signal.

[0086] When the maximum frequency shift amplitude is greater than or equal to a preset frequency shift threshold, the first ultrasonic signal is input into a pre-trained first model, and the object is detected by the first model to obtain a detection result.

[0087] Similarly, the object detection method provided by this embodiment can be applied to a variety of scenarios for detection based on object motion, and the object can be any movable object such as a person, an animal, or a vehicle. For example, when the object is a person, the object detection method can be applied to a scenario for identity authentication based on facial movements (such as shaking the head, nodding, etc.). When the object is an animal, the object detection method can be a scenario for animal species identification based on animal body movements (such as running, walking, jumping, etc.). When the object is a vehicle, the object detection method can be a scenario for vehicle performance detection based on vehicle movements (such as turning around, turning, going straight, etc.).

[0088] Of course, the above content is only an exemplary description of the object detection method, and does not limit the scene to which the object belongs. Accordingly, the type of model is also not limited. The first model may be a detection model, which is obtained by training the processed samples, and the samples may be processed by the sample processing method in any of the above embodiments.

[0089] Next, we will explain one of the authentication methods in the object detection method in detail. Figure 8is a schematic flow chart of an identity authentication method according to an embodiment of the present application, such as Figure 8 As shown, the authentication method includes the following steps: S802, performing signal quality detection on the second ultrasonic signal according to the time-frequency diagram of the second ultrasonic signal of the face to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the second ultrasonic signal within a preset time length.

[0090] S804: When the signal quality detection result satisfies a preset condition, determine a maximum frequency shift amplitude of the second ultrasonic signal within a preset time period according to the time-frequency diagram.

[0091] S806, performing face identity authentication according to the second ultrasonic signal and the maximum frequency shift amplitude to obtain a face identity authentication result.

[0092] By adopting the technical solution of the embodiment of the present application, the signal quality of the ultrasonic signal is detected according to the time-frequency diagram of the second ultrasonic signal of the face to obtain the signal quality detection result; when the signal quality detection result meets the preset conditions, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the time-frequency diagram; and then the face is authenticated according to the ultrasonic signal and the maximum frequency shift amplitude to obtain the face identity authentication result. It can be seen that when performing face identity authentication, the acquired ultrasonic signal is not directly used to perform face identity authentication, but the signal quality of the ultrasonic signal is detected in advance to ensure that the signal quality of the acquired ultrasonic signal is qualified. Moreover, when the signal quality of the ultrasonic signal is qualified, the face identity authentication is further performed according to the frequency shift amplitude of the ultrasonic signal, which improves the quality of the ultrasonic signal used for face identity authentication and ensures the accuracy of face identity authentication based on the ultrasonic signal.

[0093] In one embodiment, according to the second ultrasonic signal and the maximum frequency shift amplitude, face identity verification is performed on the face, and when the face identity verification is obtained, the following method may be performed: When the maximum frequency shift amplitude is less than the preset frequency shift threshold, it is determined that the face identity verification is detected as failed. If the maximum frequency shift amplitude is less than the preset frequency shift threshold, it means that it does not cause sufficient frequency shift and can be considered as an attack signal.

[0094] When the maximum frequency shift amplitude is greater than or equal to a preset frequency shift threshold, the second ultrasonic signal is input into a pre-trained first model, and the face is authenticated by the first model to obtain a face authentication result.

[0095] Among them, the first model can be a face recognition model or a liveness detection model, and the face recognition model or the liveness detection model is obtained by training the processed samples, and the samples can be processed by the sample processing method in any of the above embodiments.

[0096] Figure 7 and Figure 8 The examples respectively illustrate how to apply the processed ultrasonic signal (i.e., an ultrasonic signal with qualified signal quality and sufficient frequency shift) to different scenarios, including object detection scenario and identity authentication scenario, etc. In practical applications, the processed ultrasonic signal can also be applied to other scenarios, such as classification scenario, etc., which are not listed here one by one.

[0097] In summary, specific embodiments of the present subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions recorded in the claims can be performed in a different order and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing can be advantageous.

[0098] The above are the sample processing method, object detection method and identity authentication method provided in the embodiments of the present application. Based on the same idea, the embodiments of the present application also provide a sample processing device, object detection device and identity authentication device.

[0099] Fig. 9 is a schematic block diagram of a sample processing device according to an embodiment of the present application, such as Fig. 9 As shown, the device comprises: A first detection module 91 is used to perform a signal quality detection on the ultrasonic signal according to a first time-frequency diagram of the ultrasonic signal in the sample to obtain a signal quality detection result, wherein the ultrasonic signal is an ultrasonic signal extracted from the sample, and the first time-frequency diagram is used to represent the frequency shift amplitude of the ultrasonic signal within a preset time length; A first determination module 92, configured to determine, when the signal quality detection result satisfies a preset condition, a maximum frequency shift amplitude of the ultrasonic signal within the preset time length according to the first time-frequency graph; The second determination module 93 is configured to determine that the sample meets a sample screening condition when the maximum frequency shift amplitude is greater than or equal to a preset frequency shift threshold.

[0100] In one embodiment, the first detection module 91 performs the following steps when performing signal quality detection on the ultrasonic signal according to the first time-frequency diagram of the ultrasonic signal in the sample and obtaining the signal quality detection result: According to the first time-frequency graph, when the ultrasonic signal satisfies at least one of the following, it is determined that the signal quality detection result satisfies the preset condition: The average value of the spectrum energy of the ultrasonic signal in the first central offset region of the first time-frequency graph is greater than or equal to a preset energy threshold; the offset of the pixel points in the first central offset region relative to the central region of the first time-frequency graph is less than or equal to a preset offset threshold; The background noise of the ultrasonic signal meets the preset noise condition.

[0101] In one embodiment, the first detection module 91 performs the following steps when performing signal quality detection on the ultrasonic signal according to the first time-frequency diagram of the ultrasonic signal in the sample and obtaining the signal quality detection result: Binarizing the first time-frequency graph according to the spectrum energy value of each pixel point in the first time-frequency graph to obtain a second time-frequency graph; Determine a first connected domain in the second time-frequency graph; When the number of the first connected domains is less than or equal to a preset connected domain threshold, it is determined that the background noise meets the preset noise condition.

[0102] In one embodiment, when the first determination module 92 determines the maximum frequency shift amplitude of the ultrasonic signal within the preset time length according to the first time-frequency diagram, the first determination module 92 performs the following steps: Determine the maximum connected domain in the second time-frequency graph; The maximum offset of the maximum connected domain relative to the central area of ​​the second time-frequency graph is determined, and the maximum offset is determined as the maximum frequency shift amplitude.

[0103] In one embodiment, the apparatus further comprises: A fifth determination module, configured to determine the sample category of the sample ultrasound signal according to the label information carried by the sample after determining that the sample meets the sample screening condition; the sample category includes a positive sample category and a negative sample category; The training module is used to train the first model according to the samples and the sample categories.

[0104] In one embodiment, the apparatus further comprises: A discarding module is used to discard the sample when the ultrasonic signal satisfies at least one of the following conditions: The average value of the spectrum energy of the ultrasonic signal in the first center offset area is less than the preset energy threshold; The background noise of the ultrasonic signal does not meet the preset noise condition; The maximum frequency shift amplitude is smaller than the preset frequency shift threshold.

[0105] In one embodiment, the apparatus further comprises: a sixth determination module, configured to determine, before performing signal quality detection on the ultrasonic signal according to the first time-frequency graph of the ultrasonic signal in the sample, a second center offset region of the ultrasonic signal in the first time-frequency graph; the second center offset region includes the first center offset region; A shielding module, used for shielding the second center offset area to obtain a shielded time-frequency diagram; When the first detection module 91 performs signal quality detection on the ultrasonic signal according to the first time-frequency diagram of the ultrasonic signal in the sample, the first detection module 91 executes the following steps: The signal quality of the ultrasonic signal is detected according to the shielded time-frequency diagram.

[0106] Using the device of the embodiment of the present application, the signal quality detection of the ultrasonic signal is performed according to the first time-frequency diagram of the ultrasonic signal in the sample to obtain the signal quality detection result; when the signal quality detection result satisfies the preset condition, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the first time-frequency diagram; when the maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, it is determined that the sample meets the sample screening condition. It can be seen that the acquired sample is not used directly, for example, it is not directly applied to the model training scene, but the signal quality detection of the ultrasonic signal in the sample is performed in advance to ensure that the signal quality of the ultrasonic signal is qualified. In addition, when the signal quality of the ultrasonic signal is qualified, the frequency shift amplitude of the ultrasonic signal is further detected. Since a larger image area shows different modes when the frequency shift is large enough, it is beneficial for the later model to learn the difference between positive and negative samples. If the frequency shift is very small, it can be considered that the target object corresponding to the sample has no movement, which does not conform to the principle of the frequency shift phenomenon caused by the movement of the target object in this scheme. Therefore, by ensuring that the processed ultrasonic signal has sufficient frequency shift, when the sample is used, for example, when the sample is applied to the training first model scene, the quality of the sample is ensured, and the model performance of the trained first model is further improved.

[0107] Those skilled in the art should understand that Fig. 9 The sample processing device in can be used to implement the sample processing method described above, and the detailed description thereof should be similar to the description in the method part above. To avoid tediousness, it will not be repeated here.

[0108] Fig.10 is a schematic block diagram of an object detection device according to an embodiment of the present application, such as Fig.10 As shown, the device comprises: A second detection module 101 is used to perform a signal quality detection on the first ultrasonic signal of the object according to a time-frequency diagram of the first ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the first ultrasonic signal within a preset time length; A third determination module 102 is used to determine the maximum frequency shift amplitude of the first ultrasonic signal within the preset time length according to the time-frequency diagram when the signal quality detection result meets the preset condition; The third detection module 103 is used to detect the object according to the first ultrasonic signal and the maximum frequency shift amplitude to obtain a detection result of the object.

[0109] In one embodiment, when the third detection module 103 detects the object according to the first ultrasonic signal and the maximum frequency shift amplitude to obtain the detection result of the object, the following steps are performed: When the maximum frequency shift amplitude is less than a preset frequency shift threshold, determining that the detection result is a detection failure; In a case where the maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, the first ultrasonic signal is input into a pre-trained first model, and the object is detected by the first model to obtain the detection result.

[0110] By using the device of the embodiment of the present application, the signal quality detection of the ultrasonic signal is performed according to the time-frequency diagram of the first ultrasonic signal of the object to obtain the signal quality detection result; when the signal quality detection result meets the preset conditions, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the time-frequency diagram; and then the object is detected according to the ultrasonic signal and the maximum frequency shift amplitude to obtain the detection result of the object. It can be seen that when detecting the object, the acquired ultrasonic signal is not directly used for detection, but the signal quality detection of the ultrasonic signal is performed in advance to ensure that the signal quality of the acquired ultrasonic signal is qualified. Moreover, when the signal quality of the ultrasonic signal is qualified, it is further detected according to the frequency shift amplitude of the ultrasonic signal, which improves the quality of the ultrasonic signal used for detection and ensures the accuracy of detection based on the ultrasonic signal.

[0111] Those skilled in the art should understand that Fig.10 The object detection device in can be used to implement the object detection method described above, and the detailed description thereof should be similar to the description of the method part above. To avoid tediousness, it will not be repeated here.

[0112] Fig.11 is a schematic block diagram of an identity authentication device according to an embodiment of the present application, such as Fig.11 As shown, the device comprises: The fourth detection module 111 is used to perform a signal quality detection on the second ultrasonic signal of the face according to a time-frequency diagram of the second ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the second ultrasonic signal within a preset time length; A fourth determination module 112 is configured to determine, if the signal quality detection result satisfies a preset condition, a maximum frequency shift amplitude of the second ultrasonic signal within the preset time length according to the time-frequency diagram; The identity authentication module 113 is used to perform identity authentication on the face according to the second ultrasonic signal and the maximum frequency shift amplitude to obtain an identity authentication result of the face.

[0113] In one embodiment, when the identity verification module 113 performs identity verification on the face according to the second ultrasonic signal and the maximum frequency shift amplitude to obtain the identity verification result of the face, the following steps are performed: When the maximum frequency shift amplitude is less than a preset frequency shift threshold, determining that the identity authentication result of the face is detection failure; When the maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, the second ultrasonic signal is input into a pre-trained first model, and the face is authenticated by the first model to obtain an authentication result of the face.

[0114] The device of the embodiment of the present application is used to perform signal quality detection on the ultrasonic signal according to the time-frequency diagram of the second ultrasonic signal of the face to obtain a signal quality detection result; when the signal quality detection result meets the preset conditions, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the time-frequency diagram; and then the face is authenticated according to the ultrasonic signal and the maximum frequency shift amplitude to obtain an authentication result. It can be seen that when performing face identity authentication, the acquired ultrasonic signal is not directly used to perform face identity authentication, but the ultrasonic signal is pre-detected for signal quality to ensure that the signal quality of the acquired ultrasonic signal is qualified. Moreover, when the signal quality of the ultrasonic signal is qualified, the face identity authentication is further performed according to the frequency shift amplitude of the ultrasonic signal, which improves the quality of the ultrasonic signal used for face identity authentication and ensures the accuracy of face identity authentication based on the ultrasonic signal.

[0115] Those skilled in the art should understand that Fig.11 The identity authentication device in can be used to implement the identity authentication method described above, and the detailed description thereof should be similar to the description of the method part above. To avoid tediousness, it will not be repeated here.

[0116] Based on the same idea, the embodiment of the present application also provides an electronic device, such as Fig.12As shown. The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors 1201 and memory 1202, and the memory 1202 may store one or more storage applications or data. Among them, the memory 1202 may be a short-term storage or a persistent storage. The application stored in the memory 1202 may include one or more modules (not shown in the figure), and each module may include a series of computer executable instructions in the electronic device. Furthermore, the processor 1201 can be configured to communicate with the memory 1202 to execute a series of computer executable instructions in the memory 1202 on the electronic device. The electronic device may also include one or more power supplies 1203, one or more wired or wireless network interfaces 1204, one or more input and output interfaces 1205, and one or more keyboards 1206.

[0117] Specifically in this embodiment, the electronic device includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions for the electronic device, and the one or more programs are configured to be executed by one or more processors, including the following computer executable instructions: According to a first time-frequency graph of an ultrasonic signal in a sample, a signal quality detection is performed on the ultrasonic signal to obtain a signal quality detection result, wherein the ultrasonic signal is an ultrasonic signal extracted from the sample, and the first time-frequency graph is used to represent a frequency shift amplitude of the ultrasonic signal within a preset time length; When the signal quality detection result satisfies a preset condition, determining a maximum frequency shift amplitude of the ultrasonic signal within the preset time length according to the first time-frequency graph; When the maximum frequency shift amplitude is greater than or equal to a preset frequency shift threshold, it is determined that the sample meets the sample screening condition.

[0118] According to the technical solution of the embodiment of the present application, the signal quality detection of the ultrasonic signal is performed according to the first time-frequency diagram of the ultrasonic signal in the sample to obtain the signal quality detection result; when the signal quality detection result satisfies the preset condition, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the first time-frequency diagram; when the maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, it is determined that the sample meets the sample screening condition. It can be seen that the acquired sample is not used directly, for example, it is not directly applied to the model training scene, but the signal quality detection of the ultrasonic signal in the sample is performed in advance to ensure that the signal quality of the ultrasonic signal is qualified. In addition, when the signal quality of the ultrasonic signal is qualified, the frequency shift amplitude of the ultrasonic signal is further detected. Since a large image area shows different modes when the frequency shift is large enough, it is beneficial for the later model to learn the difference between positive and negative samples. If the frequency shift is very small, it can be considered that the target object corresponding to the sample has no movement, which does not conform to the principle of the frequency shift phenomenon caused by the movement of the target object in this solution. Therefore, by ensuring that the processed ultrasonic signal has sufficient frequency shift, when the sample is used, for example, when the sample is applied to the training liveness detection model scene, the quality of the sample is ensured, and the model performance of the trained first model is further improved.

[0119] Specifically in another embodiment, the electronic device includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions in the electronic device, and the one or more programs are configured to be executed by one or more processors, including computer executable instructions for performing the following: According to a time-frequency diagram of a first ultrasonic signal of the object, a signal quality detection is performed on the first ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the first ultrasonic signal within a preset time length; In the case where the signal quality detection result satisfies the preset condition, determining the maximum frequency shift amplitude of the first ultrasonic signal within the preset time length according to the time-frequency diagram; The object is detected according to the first ultrasonic signal and the maximum frequency shift amplitude to obtain a detection result of the object.

[0120] By adopting the technical solution of the embodiment of the present application, the signal quality detection of the ultrasonic signal is performed according to the time-frequency diagram of the first ultrasonic signal of the object to obtain the signal quality detection result; when the signal quality detection result meets the preset conditions, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the time-frequency diagram; and then the object is detected according to the ultrasonic signal and the maximum frequency shift amplitude to obtain the detection result of the object. It can be seen that when detecting the object, the acquired ultrasonic signal is not directly used for detection, but the signal quality detection of the ultrasonic signal is performed in advance to ensure that the signal quality of the acquired ultrasonic signal is qualified. Moreover, when the signal quality of the ultrasonic signal is qualified, it is further detected according to the frequency shift amplitude of the ultrasonic signal, which improves the quality of the ultrasonic signal used for detection and ensures the accuracy of detection based on the ultrasonic signal.

[0121] Specifically in another embodiment, the electronic device includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions in the electronic device, and the one or more programs are configured to be executed by one or more processors, including computer executable instructions for performing the following: According to the time-frequency diagram of the second ultrasonic signal of the human face, a signal quality detection is performed on the second ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the second ultrasonic signal within a preset time length; In the case where the signal quality detection result satisfies the preset condition, determining the maximum frequency shift amplitude of the second ultrasonic signal within the preset time length according to the time-frequency diagram; The human face is authenticated according to the second ultrasonic signal and the maximum frequency shift amplitude to obtain an identity authentication result of the human face.

[0122] By adopting the technical solution of the embodiment of the present application, the signal quality of the ultrasonic signal is detected according to the time-frequency diagram of the second ultrasonic signal of the face to obtain the signal quality detection result; when the signal quality detection result meets the preset conditions, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the time-frequency diagram; and then the face is authenticated according to the ultrasonic signal and the maximum frequency shift amplitude to obtain the identity authentication result. It can be seen that when performing identity authentication, the ultrasonic signal obtained is not directly used for identity authentication, but the ultrasonic signal is pre-detected for signal quality to ensure that the signal quality of the obtained ultrasonic signal is qualified. Moreover, when the signal quality of the ultrasonic signal is qualified, the identity authentication is further performed according to the frequency shift amplitude of the ultrasonic signal, which improves the quality of the ultrasonic signal used for identity authentication and ensures the accuracy of identity authentication based on the ultrasonic signal.

[0123] The embodiment of the present application further proposes a computer-readable storage medium, which stores one or more computer programs, and the one or more computer programs include instructions. When the instructions are executed by an electronic device including multiple application programs, the electronic device can perform each process of the above sample processing method embodiment, and is specifically used to perform: According to a first time-frequency graph of an ultrasonic signal in a sample, a signal quality detection is performed on the ultrasonic signal to obtain a signal quality detection result, wherein the ultrasonic signal is an ultrasonic signal extracted from the sample, and the first time-frequency graph is used to represent a frequency shift amplitude of the ultrasonic signal within a preset time length; When the signal quality detection result satisfies a preset condition, determining a maximum frequency shift amplitude of the ultrasonic signal within the preset time length according to the first time-frequency graph; When the maximum frequency shift amplitude is greater than or equal to a preset frequency shift threshold, it is determined that the sample meets the sample screening condition.

[0124] According to the technical solution of the embodiment of the present application, the signal quality detection of the ultrasonic signal is performed according to the first time-frequency diagram of the ultrasonic signal in the sample to obtain the signal quality detection result; when the signal quality detection result satisfies the preset condition, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the first time-frequency diagram; when the maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, it is determined that the sample meets the sample screening condition. It can be seen that the acquired sample is not used directly, for example, it is not directly applied to the model training scene, but the signal quality detection of the ultrasonic signal in the sample is performed in advance to ensure that the signal quality of the ultrasonic signal is qualified. In addition, when the signal quality of the ultrasonic signal is qualified, the frequency shift amplitude of the ultrasonic signal is further detected. Since a large image area shows different modes when the frequency shift is large enough, it is beneficial for the later model to learn the difference between positive and negative samples. If the frequency shift is very small, it can be considered that the target object corresponding to the sample has no movement, which does not conform to the principle of the frequency shift phenomenon caused by the movement of the target object in this solution. Therefore, by ensuring that the processed ultrasonic signal has sufficient frequency shift, when the sample is used, for example, when the sample is applied to the training first model scene, the quality of the sample is ensured, and the model performance of the trained first model is further improved.

[0125] The embodiment of the present application further proposes a computer-readable storage medium, which stores one or more computer programs. The one or more computer programs include instructions. When the instructions are executed by an electronic device including multiple application programs, the electronic device can perform various processes of the above-mentioned object detection method embodiment, and are specifically used to perform: According to a time-frequency diagram of a first ultrasonic signal of the object, a signal quality detection is performed on the first ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the first ultrasonic signal within a preset time length; In the case where the signal quality detection result satisfies the preset condition, determining the maximum frequency shift amplitude of the first ultrasonic signal within the preset time length according to the time-frequency diagram; The object is detected according to the first ultrasonic signal and the maximum frequency shift amplitude to obtain a detection result of the object.

[0126] By adopting the technical solution of the embodiment of the present application, the signal quality detection of the ultrasonic signal is performed according to the time-frequency diagram of the first ultrasonic signal of the object to obtain the signal quality detection result; when the signal quality detection result meets the preset conditions, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the time-frequency diagram; and then the object is detected according to the ultrasonic signal and the maximum frequency shift amplitude to obtain the detection result of the object. It can be seen that when performing liveness detection on the object, the acquired ultrasonic signal is not directly used for liveness detection, but the signal quality detection of the ultrasonic signal is performed in advance to ensure that the signal quality of the acquired ultrasonic signal is qualified. Moreover, when the signal quality of the ultrasonic signal is qualified, it is further detected according to the frequency shift amplitude of the ultrasonic signal, which improves the quality of the ultrasonic signal used for liveness detection and ensures the accuracy of detection based on the ultrasonic signal.

[0127] The present application also provides a computer-readable storage medium, which stores one or more computer programs. The one or more computer programs include instructions. When the instructions are executed by an electronic device including multiple application programs, the electronic device can perform various processes of the above-mentioned identity authentication method embodiment, and are specifically used to perform: According to the time-frequency diagram of the second ultrasonic signal of the human face, a signal quality detection is performed on the second ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the second ultrasonic signal within a preset time length; In the case where the signal quality detection result satisfies the preset condition, determining the maximum frequency shift amplitude of the second ultrasonic signal within the preset time length according to the time-frequency diagram; The human face is authenticated according to the second ultrasonic signal and the maximum frequency shift amplitude to obtain an identity authentication result of the human face.

[0128] By adopting the technical solution of the embodiment of the present application, the signal quality of the ultrasonic signal is detected according to the time-frequency diagram of the second ultrasonic signal of the face to obtain the signal quality detection result; when the signal quality detection result meets the preset conditions, the maximum frequency shift amplitude of the ultrasonic signal within the preset time length is determined according to the time-frequency diagram; and then the face is authenticated according to the ultrasonic signal and the maximum frequency shift amplitude to obtain the identity authentication result. It can be seen that when performing identity authentication, the ultrasonic signal obtained is not directly used for identity authentication, but the ultrasonic signal is pre-detected for signal quality to ensure that the signal quality of the obtained ultrasonic signal is qualified. Moreover, when the signal quality of the ultrasonic signal is qualified, the identity authentication is further performed according to the frequency shift amplitude of the ultrasonic signal, which improves the quality of the ultrasonic signal used for identity authentication and ensures the accuracy of identity authentication based on the ultrasonic signal.

[0129] An embodiment of the present application provides a computer program product, including a computer program, wherein the computer program is executed by a processor to implement the various processes of the sample processing method embodiment as described above, or the computer program is executed by a processor to implement the various processes of the object detection method embodiment as described above, or the computer program is executed by a processor to implement the various processes of the identity authentication method embodiment as described above, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0130] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0131] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0132] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0133] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0134] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0135] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0136] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0137] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0138] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined in this article, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0139] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0140] The present application may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0141] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0142] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A sample processing method, characterized in that: include: According to a first time-frequency graph of an ultrasonic signal in a sample, a signal quality detection is performed on the ultrasonic signal to obtain a signal quality detection result, wherein the ultrasonic signal is an ultrasonic signal extracted from the sample, and the first time-frequency graph is used to represent a frequency shift amplitude of the ultrasonic signal within a preset time length; When the signal quality detection result satisfies a preset condition, determining a maximum frequency shift amplitude of the ultrasonic signal within the preset time length according to the first time-frequency graph; When the maximum frequency shift amplitude is greater than or equal to a preset frequency shift threshold, it is determined that the sample meets the sample screening condition.

2. The method according to claim 1, characterized in that The step of performing signal quality detection on the ultrasonic signal according to the first time-frequency diagram of the ultrasonic signal in the sample to obtain a signal quality detection result includes: According to the first time-frequency graph, when the ultrasonic signal satisfies at least one of the following, it is determined that the signal quality detection result satisfies the preset condition: The average value of the spectrum energy of the ultrasonic signal in the first center offset region of the first time-frequency diagram is greater than or equal to a preset energy threshold; The offset of the pixel points in the first central offset area relative to the central area of ​​the first time-frequency graph is less than or equal to a preset offset threshold; The background noise of the ultrasonic signal meets the preset noise condition.

3. The method according to claim 2, characterized in that The step of performing signal quality detection on the ultrasonic signal according to the first time-frequency diagram of the ultrasonic signal in the sample to obtain a signal quality detection result includes: Binarizing the first time-frequency graph according to the spectrum energy value of each pixel point in the first time-frequency graph to obtain a second time-frequency graph; Determine a first connected domain in the second time-frequency graph; When the number of the first connected domains is less than or equal to a preset connected domain threshold, it is determined that the background noise meets the preset noise condition.

4. The method according to claim 3, characterized in that The step of determining the maximum frequency shift amplitude of the ultrasonic signal within the preset time period according to the first time-frequency graph includes: Determine the maximum connected domain in the second time-frequency graph; The maximum offset between the maximum connected domain and the central area of ​​the second time-frequency graph is calculated, and the maximum offset is determined as the maximum frequency shift amplitude.

5. The method according to any one of claims 1 to 4, characterized in that: After determining that the sample meets the sample screening condition, the method further includes: Determine the sample category of the sample according to the label information carried by the sample; the sample category includes a positive sample category and a negative sample category; The first model is trained according to the samples and the sample categories.

6. The method according to claim 2, characterized in that The method further comprises: If the ultrasound signal satisfies at least one of the following conditions, the sample is discarded: The average value of the spectrum energy of the ultrasonic signal in the first center offset area is less than the preset energy threshold; The background noise of the ultrasonic signal does not meet the preset noise condition; The maximum frequency shift amplitude is smaller than the preset frequency shift threshold.

7. The method according to claim 2, characterized in that: Before performing signal quality detection on the ultrasonic signal according to the first time-frequency diagram of the ultrasonic signal in the sample, the method further includes: Determine a second central offset region of the ultrasonic signal in the first time-frequency diagram; the second central offset region includes the first central offset region; Performing shielding processing on the second center offset region to obtain a shielded time-frequency diagram; The step of performing signal quality detection on the ultrasonic signal according to the first time-frequency diagram of the ultrasonic signal in the sample comprises: The signal quality of the ultrasonic signal is detected according to the shielded time-frequency diagram.

8. An object detection method, characterized in that: include: According to a time-frequency diagram of a first ultrasonic signal of the object, a signal quality detection is performed on the first ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the first ultrasonic signal within a preset time length; In the case where the signal quality detection result satisfies the preset condition, determining the maximum frequency shift amplitude of the first ultrasonic signal within the preset time length according to the time-frequency diagram; The object is detected according to the first ultrasonic signal and the maximum frequency shift amplitude to obtain a detection result of the object.

9. The method according to claim 8, characterized in that The detecting the object according to the first ultrasonic signal and the maximum frequency shift amplitude to obtain a detection result of the object includes: When the maximum frequency shift amplitude is less than a preset frequency shift threshold, determining that the detection result is a detection failure; In a case where the maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, the first ultrasonic signal is input into a pre-trained first model, and the object is detected by the first model to obtain the detection result.

10. An identity authentication method, characterized in that: include: According to the time-frequency diagram of the second ultrasonic signal of the human face, a signal quality detection is performed on the second ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the second ultrasonic signal within a preset time length; In the case where the signal quality detection result satisfies the preset condition, determining the maximum frequency shift amplitude of the second ultrasonic signal within the preset time length according to the time-frequency diagram; The human face is authenticated according to the second ultrasonic signal and the maximum frequency shift amplitude to obtain an identity authentication result of the human face.

11. The method according to claim 10, characterized in that The step of performing identity authentication on the face according to the second ultrasonic signal and the maximum frequency shift amplitude to obtain an identity authentication result of the face includes: When the maximum frequency shift amplitude is less than a preset frequency shift threshold, determining that the identity authentication result is a detection failure; When the maximum frequency shift amplitude is greater than or equal to the preset frequency shift threshold, the second ultrasonic signal is input into a pre-trained first model, and the face identity is authenticated by the first model to obtain the identity authentication result.

12. A sample processing device, characterized in that: include: a first detection module, configured to perform a signal quality detection on the ultrasonic signal according to a first time-frequency graph of the ultrasonic signal in the sample to obtain a signal quality detection result, wherein the ultrasonic signal is an ultrasonic signal extracted from the sample, and the first time-frequency graph is used to represent a frequency shift amplitude of the ultrasonic signal within a preset time length; A first determination module, configured to determine, when the signal quality detection result satisfies a preset condition, a maximum frequency shift amplitude of the ultrasonic signal within the preset time length according to the first time-frequency graph; The second determination module is used to determine that the sample meets the sample screening condition when the maximum frequency shift amplitude is greater than or equal to a preset frequency shift threshold.

13. An object detection device, characterized in that: include: A second detection module is used to perform a signal quality detection on the first ultrasonic signal of the object according to a time-frequency diagram of the first ultrasonic signal to obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the first ultrasonic signal within a preset time length; A third determination module is used to determine, when the signal quality detection result satisfies a preset condition, a maximum frequency shift amplitude of the first ultrasonic signal within the preset time length according to the time-frequency diagram; The third detection module is used to detect the object according to the first ultrasonic signal and the maximum frequency shift amplitude to obtain a detection result of the object.

14. An identity verification device, characterized in that: include: a fourth detection module, configured to perform a signal quality detection on the second ultrasonic signal of the face according to a time-frequency diagram of the second ultrasonic signal, and obtain a signal quality detection result; the time-frequency diagram is used to represent the frequency shift amplitude of the second ultrasonic signal within a preset time length; a fourth determination module, configured to determine, if the signal quality detection result satisfies a preset condition, a maximum frequency shift amplitude of the second ultrasonic signal within the preset time length according to the time-frequency diagram; An identity authentication module is used to authenticate the face according to the second ultrasonic signal and the maximum frequency shift amplitude to obtain an identity authentication result of the face.

15. An electronic device, characterized in that: It includes a processor and a memory electrically connected to the processor, the memory storing a computer program, the processor being used to call and execute the computer program from the memory to implement the sample processing method as described in any one of claims 1 to 7, or the processor being used to call and execute the computer program from the memory to implement the object detection method as described in any one of claims 8 to 9, or the processor being used to call and execute the computer program from the memory to implement the identity authentication method as described in any one of claims 10 to 11.

16. A computer-readable storage medium, characterized in that: The storage medium is used to store a computer program, which can be executed by a processor to implement the sample processing method as described in any one of claims 1-7, or, the storage medium is used to store a computer program, which can be executed by a processor to implement the object detection method as described in any one of claims 8-9, or, the storage medium is used to store a computer program, which can be executed by a processor to implement the identity authentication method as described in any one of claims 10-11.

17. A computer program product, characterized in that It comprises a computer program, which is executed by a processor to implement the sample processing method as described in any one of claims 1 to 7, or the computer program is executed by a processor to implement the object detection method as described in any one of claims 8 to 9, or the computer program is executed by a processor to implement the identity authentication method as described in any one of claims 10 to 11.

Images