WO2022205643A1

WO2022205643A1 - Living body detection method and apparatus, and device and computer storage medium

Info

Publication number: WO2022205643A1
Application number: PCT/CN2021/103167
Authority: WO
Inventors: 方倩; 胥鹏; 张捷
Original assignee: 上海商汤智能科技有限公司
Priority date: 2021-03-31
Filing date: 2021-06-29
Publication date: 2022-10-06
Also published as: CN113095180A

Abstract

The present disclosure relates to a living body detection method and apparatus, and a device and a computer storage medium. The living body detection method comprises: determining an ith image under test from video data collected by an image sensor; and in response to the ith image under test satisfying a preset quality condition, performing living body detection processing on the ith image under test to obtain a target detection result. By means of the present disclosure, high-efficiency living body detection is realized.

Description

Living body detection method and device, equipment, computer storage medium

CROSS-REFERENCE TO RELATED APPLICATIONS

The present disclosure is based on the Chinese patent application with the application number of 202110346603.7, the application date of March 31, 2021, and the application title of "living detection method and device, equipment, computer storage medium", and claims the priority of the Chinese patent application, The entire contents of this Chinese patent application are hereby incorporated by reference into the present disclosure.

technical field

The present disclosure relates to the technical field of living body detection, and in particular, to a living body detection method, apparatus, device, and computer storage medium.

Background technique

With the rapid development of face recognition technology, more and more face recognition systems are used in security. As a crucial item in face recognition technology, "face brushing" occupies an important position in the security of the entire face recognition system due to its convenient operation, rapidity and high discrimination of image authenticity. Such as face check-in, face payment, face unlock and face authentication.

Since attackers often use static photos, face models, or face masks to conduct security intrusion attacks, the system first needs to perform liveness detection during the "face brushing" process, that is, it needs to confirm whether the verifier is a legitimate creature. Living body, whether it is a living, on-site, real person.

However, the related art generally has the problems of large network bandwidth occupation and long detection period when performing liveness detection, resulting in low detection efficiency.

SUMMARY OF THE INVENTION

Embodiments of the present disclosure provide a method, apparatus, device, and computer storage medium for detecting a living body.

The technical solution of the present disclosure is realized as follows:

Embodiments of the present disclosure provide a method for detecting a living body, including:

Obtain the i-th image to be inspected from the video data collected by the image sensor; wherein, i is an integer greater than 1; in response to the i-th image to be inspected meeting a preset quality condition, perform live detection on the i-th image to be inspected process to obtain the target detection result.

In the above method, the acquiring the i-th image to be inspected from the video data collected by the image sensor includes: extracting an initial image from the video data collected by the image sensor; pre-formatting the initial image according to preset configuration parameters. processing to obtain a processed image; and determining the processed image as the i-th image to be inspected.

An embodiment of the present disclosure provides a device for detecting a living body, including an acquisition module and a first processing module,

The acquisition module is configured to acquire the i-th image to be inspected from the video data collected by the image sensor; wherein, the i is an integer greater than 1;

The first processing module is configured to, in response to the i-th image to be inspected meeting a preset quality condition, perform in vivo detection processing on the i-th image to be inspected to obtain a target detection result.

An embodiment of the present disclosure provides a living body detection device, including: a processor and a memory storing instructions executable by the processor, and when the instructions are executed by the processor, the above-mentioned living body detection method is implemented.

An embodiment of the present disclosure provides a computer-readable storage medium storing a program, and when the program is executed by a processor, the above-mentioned method for detecting a living body is implemented.

An embodiment of the present disclosure provides a computer program, including computer-readable codes, and when the computer-readable codes are executed in an electronic device and executed by a processor in the electronic device, the above-mentioned living body detection is realized method.

According to the technical solution provided by the embodiments of the present disclosure, the living body detection device first obtains the i-th image to be inspected from the video data collected by the image sensor; and then responds that the i-th image to be inspected meets the preset quality condition, Liveness detection processing is performed to obtain target detection results. In this way, the detection process is based on continuous images in the process of living body detection. Compared with video living body detection, the network bandwidth is less occupied, and timely feedback of the detection results can be realized, and the quality detection process is added before the subsequent living body detection, which effectively ensures the detection. data quality in the process. It can be seen that the living body detection solution proposed by the embodiments of the present disclosure effectively solves the problems of large network bandwidth occupation and long detection period, and realizes high-efficiency living body detection.

Description of drawings

FIG. 1 is a schematic diagram of the structure of a living body detection method in the related art;

FIG. 2 is a schematic diagram 1 of the implementation flow of the live detection method proposed by the embodiment of the present disclosure;

FIG. 3 is a schematic diagram 2 of the implementation flow of the method for detecting a living body provided in an embodiment of the present disclosure;

FIG. 4 is a schematic diagram 3 of the implementation flow of the method for detecting a living body proposed by an embodiment of the present disclosure;

FIG. 5 is a fourth schematic diagram of the implementation flow of the method for detecting a living body provided by an embodiment of the present disclosure;

FIG. 6 is a schematic diagram 5 of the implementation flow of the living body detection method provided by the embodiment of the present disclosure;

FIG. 7 is a sixth schematic diagram of the implementation process of the living body detection method provided by the embodiment of the present disclosure;

FIG. 8 is a schematic diagram 7 of the implementation flow of the method for detecting a living body provided by an embodiment of the present disclosure;

FIG. 9 is a schematic diagram 8 of the implementation flow of the living body detection method provided by the embodiment of the present disclosure;

FIG. 10 is a schematic diagram 9 of the implementation flow of the living body detection method provided by the embodiment of the present disclosure;

FIG. 11 is a schematic diagram ten of the implementation flow of the method for detecting a living body provided by an embodiment of the present disclosure;

FIG. 12 is a schematic diagram of an application scenario of the living body detection method proposed by the embodiment of the present disclosure;

FIG. 13 is a schematic diagram of the execution flow of the live detection method proposed by the embodiment of the present disclosure;

FIG. 14 is a schematic diagram of the composition and structure of a living body detection device proposed by an embodiment of the present disclosure;

FIG. 15 is a schematic diagram of the composition and structure of a living body detection device provided by an embodiment of the present disclosure.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present disclosure clearer, the present disclosure will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations of the present disclosure, and those skilled in the art will not All other embodiments obtained under the premise of creative work fall within the protection scope of the present disclosure.

In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is understood that "some embodiments" can be the same or a different subset of all possible embodiments, and Can be combined with each other without conflict.

In the following description, the term "first\second\third" is only used to distinguish similar objects, and does not represent a specific ordering of objects. It is understood that "first\second\third" Where permitted, the specific order or sequence may be interchanged to enable the embodiments of the disclosure described herein to be practiced in sequences other than those illustrated or described herein.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. The terminology used herein is for the purpose of describing the embodiments of the present disclosure only and is not intended to limit the present disclosure.

At present, "face brushing", as a crucial item in face recognition technology, occupies an important position in the security of the entire face recognition system due to its convenient operation, rapidity and high discrimination of image authenticity. Such as face check-in, face payment, face unlock and face authentication.

When users use various hypertext (HyperText Markup Language 5, H5) pages to run on the browser, identity authentication is required, such as identity authentication through face recognition. In face recognition, attackers often use static photos, face models or face masks to conduct security intrusion attacks. Therefore, in the process of "swiping face", the system first needs to perform live detection, that is, it needs to confirm the verification first. Whether the person is a legitimate living organism, whether it is a living, on-site, real person.

However, when performing live detection in the related art, a common method is that the front end of the system calls a camera to capture video and transmit it to the back end, then the back end receives the video and performs live body detection, and after the video detection is completed, the result is fed back to the front end. For example, Figure 1 is a schematic diagram of the architecture of a living body detection method in the related art. As shown in Figure 1, the front end of the system collects video data by calling a camera, and then transmits the collected video data to the back end. The back end uses a living body detection module, based on The relevant algorithm performs live detection on the received video data, and after obtaining the detection result, the living body detection result is fed back to the front end.

It can be understood that the living body detection scheme in the related art has the following defects: firstly, video data is used. Since the video data is generally between 5 and 20M, there is a problem that the network bandwidth is occupied; secondly, the back-end The detection results can be fed back only after the video data is subjected to silent live video detection. The duration is generally about 3 to 5 seconds, and the detection period is relatively long. In the third aspect, the data quality during the detection process cannot be guaranteed, which may affect the duration of the entire detection period.

To sum up, the living video detection in the related art generally has the problems of large network bandwidth occupation and long detection period, resulting in low detection efficiency. In view of this, the content to be discussed in the embodiments of the present disclosure how to achieve high-efficiency detection of living bodies will be described below with reference to the following specific embodiments.

Embodiments of the present disclosure provide a method, device, device, and computer storage medium for living body detection. In the living body detection process, detection is performed based on continuous images. Compared with video living body detection, the network bandwidth is occupied less, and the detection result can be timely. Feedback, and adding a quality detection process before subsequent live detection, effectively guarantees the data quality in the detection process. It can be seen that the living body detection solution proposed by the embodiments of the present disclosure effectively solves the problems of large network bandwidth occupation and long detection period, and realizes high-efficiency living body detection.

The living body detection method provided by the embodiments of the present disclosure is applied to a living body detection device. Exemplary applications of the liveness detection device provided by the embodiments of the present disclosure are described below. The liveness detection device provided by the embodiments of the present disclosure can be implemented as mobile phones, notebook computers, tablet computers, desktop computers, smart TVs, in-vehicle devices, wearable devices, industrial equipment, etc.

Below, the technical solutions in the embodiments of the present disclosure will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present disclosure.

An embodiment of the present disclosure provides a method for detecting a living body. FIG. 2 is a schematic diagram 1 of the implementation flow of the method for detecting a living body proposed in an embodiment of the present disclosure. As shown in FIG. 2 , in the embodiment of the present disclosure, a living body detection device executes The method of detection may include the following steps:

S101. Acquire an i-th image to be inspected from video data collected by an image sensor; wherein, i is an integer greater than 1.

In the embodiment of the present disclosure, the living body detection device may first acquire the i-th image to be inspected from the video data collected by the image sensor.

In some embodiments, the image sensor is an image acquisition device, that is, a camera, configured on the living body detection device. For example, in the case where the living body detection device is a smartphone, the smartphone includes a front camera and a rear camera, and the image sensor can be the front camera; when the living body detection device is a laptop, the image sensor is the The front camera of the computer.

Correspondingly, the video data is the video stream of the detection object collected by the living body detection device using the above-mentioned image sensor.

In some embodiments, the living body detection device may trigger the activation of the camera based on the user's touch operation, and then collect video data in real time through the camera.

For example, the identity verification process when opening a bank account on a mobile phone. Here, the mobile phone can receive the SMS reminder sent by the bank to the customer to open an account through the unified SMS marketing tool. The page on the mobile phone used to display the content of the SMS message is an H5 page. The external link to the bank's official website, that is, H5 jumps to the external link. After the user clicks on the external link, the mobile phone detects the trigger operation input by the user, controls the camera to turn on and collects image data in real time, and obtains the video stream.

In some embodiments, the image to be inspected refers to a picture frame to be inspected. Here, the living body detection device may select a picture frame from the video stream obtained by collection; wherein, the living body detection device may select a picture frame according to a preset time interval. Wherein, the i-th image to be inspected refers to the picture frame selected from the video stream for the i-th time during the current live detection.

For example, the living body detection device sets a timer, the timer duration is 100ms, and the living body detection device extracts a picture frame from the video stream collected by the camera every 100ms as the image to be checked.

It can be seen that the living body detection device no longer performs live body detection on the video data for the detection object, but selects images to be inspected from the video data according to time intervals, and then performs live body detection based on the images to be inspected, which reduces the occupation of network bandwidth.

S102. In response to the i-th image to be inspected meeting a preset quality condition, perform a living body detection process on the i-th image to be inspected to obtain a target detection result.

In the embodiment of the present disclosure, after the living body detection device selects the i-th image to be inspected from the video data collected by the image sensor, the living body detection device may determine that the i-th image to be inspected satisfies the preset quality condition, Liveness detection is performed on the i-th image to be inspected.

In some embodiments, the preset quality condition refers to a quality requirement for an image to be inspected that is preset by the living body detection device, such as a requirement for a detection object, a requirement for the quality of a human face, and the like.

Here, after selecting the ith image to be inspected, the living body detection device may first perform a preliminary screening based on quality requirements on the image to be inspected, so as to determine an image to be inspected whose quality meets the standard.

Wherein, the living body detection device can first determine the time parameter corresponding to the i-th image to be inspected, and judge whether the image to be inspected satisfies the initial detection condition based on the time parameter; if it is determined that the initial detection condition is met, the living body detection device can continue to target The image to be inspected is subjected to face detection, and in the case that the face detection result is that there are facial features, the quality detection value of the image is determined according to the image parameters of the image to be inspected, and then based on the quality detection value, the image to be inspected is determined whether the i-th image meets the requirements. Preset quality conditions for judgment. Further, in the case of determining that the image to be inspected satisfies the preset quality condition, a living body detection process is performed, and a detection result is obtained.

The living body detection processing refers to detecting whether there is a legitimate living body in the i-th image to be checked. The legal living body here refers to an individual with real life characteristics in the natural environment, rather than non-living characteristics such as photos or videos.

In some embodiments, the living body detection device may integrate a living body detection network, and in the case of determining that the i-th image to be inspected satisfies the preset quality conditions, call the living body detection network to judge whether there is a legitimate living body in the i-th image to be inspected , to obtain the liveness detection result of the i-th image to be inspected.

Wherein, the living body detection result of the i-th image to be inspected may be that there is a living body; it may also be that there is no living body.

In one implementation of the embodiments of the present disclosure, the living body detection device may store the living body detection result of each image to be inspected, and perform a number accumulation process in the case where the living body detection result is the presence of a living body to obtain a number accumulated value. After the in vivo detection result of the i-th image to be inspected is determined, the in vivo detection result may update the current cumulative value based on the in vivo detection result of the i-th image to be inspected, and then determine the above-mentioned target detection result based on the updated cumulative value.

Here, the target detection result may be that the in vivo verification is passed, or it may be that the in vivo verification is not passed.

It can be seen that the live detection method provided by the embodiments of the present disclosure performs detection based on continuous images during the live detection process. Compared with video live detection, the network bandwidth is occupied less, and timely feedback of the detection results can be realized. The quality detection process is added before the live detection, which effectively ensures the data quality in the detection process. It can be seen that the living body detection solution proposed by the embodiments of the present disclosure effectively solves the problems of large network bandwidth occupation and long detection period, and realizes high-efficiency living body detection.

It should be noted that various types of networks referred to in the embodiments of the present disclosure can be implemented by using neural networks, but a model composed of multiple neural networks or including multiple sub-networks is not limited herein.

Based on the above embodiment, in yet another embodiment of the present disclosure, FIG. 3 is a second schematic diagram of the implementation flow of the living body detection method provided by the embodiment of the present disclosure. As shown in FIG. 3 , in the embodiment of the present disclosure, the living body detection device The method for obtaining the i-th image to be inspected from the video data collected by the image sensor may include the following steps:

S101a, extracting an initial image from video data collected by an image sensor.

S101b. Perform format preprocessing on the initial image according to preset configuration parameters to obtain a processed image.

S101c, determining the processed image as the i-th image to be inspected.

In this embodiment of the present disclosure, in the process of selecting an image to be inspected, the living body detection device may select an image to be inspected from images that have undergone format preprocessing.

In some embodiments, since the image format input to the living body detection network may have specific requirements on the amount of image data and the size of the image when performing living body detection, the living body detection device may first select from the collected video data according to time intervals The initial image is then subjected to format conversion processing according to the actual detection requirements, so that the processed image that meets the detection requirements is determined as the image to be inspected.

For example, the mobile phone selects an image to be inspected every 100ms. Among them, after the mobile phone selects the initial image from the video stream collected by the camera of the mobile phone for the third time at 300 ms, the initial image can be resized, and the resized image can be compressed according to the preset compression ratio. , so that the compressed image is determined as the third image to be inspected.

It can be seen that, in the living body detection method provided by the embodiment of the present disclosure, the living body detection device first selects an initial image from the video data collected by the image sensor, then performs format preprocessing, and determines the image obtained after processing as the i-th image to be inspected . In this way, the custom configuration of images can be realized in the process of living body detection to meet the detection requirements.

Based on the above-mentioned embodiment, in yet another embodiment of the present disclosure, FIG. 4 is a schematic diagram 3 of the implementation flow of the living body detection method proposed by the embodiment of the present disclosure. As shown in FIG. After the i-th image to be inspected is acquired in , i.e. after step 101, and in response to the i-th image to be inspected meeting the preset quality condition, the in vivo detection process is performed on the i-th image to be inspected, and before the target detection result is obtained, that is, before step 102, The method also includes the following steps:

S103: Determine a time parameter corresponding to the i-th image to be inspected, and determine whether the i-th image to be inspected satisfies the initial detection condition according to the time parameter.

In the embodiment of the present disclosure, in order to avoid the life detection period being too long and the power consumption of the device being wasted, the living body detection device can set an initial detection condition that meets the detection requirements, and then judge whether the detection times out and whether the image is repeated based on the initial detection condition. .

In some embodiments, after obtaining the i-th image to be inspected, the living body detection device may first determine a time parameter corresponding to the image to be inspected, and determine whether the time parameter satisfies the initial detection condition.

The time parameter refers to the time difference between the timestamp of the first image to be inspected and the timestamp of the current image to be inspected, that is, the detection duration from the start of detection to the present. Here, the time parameter may be the time interval between the first image to be inspected and the i-th image to be inspected.

In an implementation of the embodiment of the present disclosure, the initial detection condition may include a detection timeout duration threshold preset by the living body detection device, and the detection is determined based on the time interval between the first to-be-detected image and the current i-th to-be-detected image. After starting the detection duration so far, the living body detection device may compare the detection duration with a preset duration threshold, and then judge whether the initial detection condition is currently satisfied based on the comparison result.

In another implementation of the embodiment of the present disclosure, the initial detection condition may further include an image repetition condition preset by the living body detection device, and the living body detection device determines the image repetition condition based on the time interval from the first to-be-detected image to the current i-th to-be-detected image. After the detection time since the detection started, the image repetition judgment can also be performed based on the image repetition condition, that is, the similarity between the current i-th image to be inspected and the historical multi-frame images to be inspected so far is compared, and then combined with the comparison result of the detection time length and The image similarity comparison result judges whether the current initial detection condition is satisfied.

In some embodiments of the present disclosure, FIG. 5 is a fourth schematic diagram of the implementation flow of the living body detection method provided by the embodiment of the present disclosure. As shown in FIG. 5 , the living body detection device is in the process between the first to-be-checked image to the i-th to-be-checked image. After the time interval is determined as the time parameter (S103a), the method for judging whether the i-th image to be inspected satisfies the initial detection condition according to the time parameter may include the following steps:

S103b1, if the time parameter is less than or equal to the preset duration threshold, read the k historical images to be inspected corresponding to the (i-k)th to (i-1)th images to be inspected; wherein, k is an integer greater than 1 and less than i .

S103b2: If at least one image to be inspected among the k historical images to be inspected is different from the i-th image to be inspected, determine that the initial detection condition is satisfied.

In an implementation of the embodiments of the present disclosure, if the living body detection device determines the detection duration from the start of detection to the present, that is, the time interval from the first image to be inspected to the i-th image to be inspected is less than or equal to the preset detection duration threshold, it indicates that the current If the detection timeout does not occur, the living body detection device can continue to perform repeated judgment of the image to be inspected.

For example, assuming that the preset duration threshold is 10s, and the time interval between the first image to be inspected and the current i-th image to be inspected is 9s, the time interval is less than the preset duration threshold, indicating that there is currently no detection timeout.

It is understandable that legal living organisms at least have dynamic changes. Therefore, in the case that the detection object may be a legal living organism, there is a difference between the adjacent images to be inspected selected according to the time interval. Repeat the same.

In some embodiments, after performing the living body detection processing of the historical images to be inspected, the living body detection device may associate and store the historical images to be inspected and the corresponding detection results. Wherein, the storage may be performed according to the chronological order and the preset storage amount. For example, the first 10 frames of images to be inspected and their corresponding historical inspection results can be stored.

Here, when performing the repeated judgment of the image to be inspected, the living body detection device can read multiple frames of pre-stored historical images to be inspected, such as the k historical images to be inspected corresponding to the (i-k)th to (i-1)th images to be inspected. image, and compare the similarity between the current i-th image to be inspected and the k historical images to be inspected, if the obtained comparison result is that the i-th image to be inspected and at least one image to be inspected in the k historical If the same, it indicates that the detection object has dynamic changes and may be a legitimate living body. At this time, it is determined that the i-th image to be inspected satisfies the initial detection conditions.

For example, in the case where the detection does not time out, assuming that k is equal to 5, when performing repeated judgment of the image to be inspected, if the living body detection device determines that the current image to be inspected is not the same as 3 frames of images in the previous 5 frames of images to be inspected in history , that is, there is a dynamic change in the detection object, indicating that the current image to be inspected is not a duplicate image and meets the initial detection conditions.

In other embodiments of the present disclosure, FIG. 6 is a schematic diagram 5 of the implementation flow of the living body detection method provided by the embodiment of the present disclosure. As shown in FIG. 6 , the living body detection device is converting the first to-be-checked image to the i-th to-be-checked image. After the time interval between is determined as the time parameter (S103a), the method for judging whether the i-th image to be inspected satisfies the initial detection condition according to the time parameter may include the following steps:

S103b3. If the time parameter is greater than the preset duration threshold, generate a first prompt message for detecting timeout.

In one implementation of the embodiment of the present disclosure, if the living body detection device determines the detection duration from the start of detection to the present, that is, the time interval from the first to-be-detected image to the i-th to-be-detected image is greater than the preset detection duration threshold, it indicates that a detection timeout currently occurs Happening.

For example, assuming that the preset duration threshold is 10s, when the time interval between the first image to be inspected and the current i-th image to be inspected is 12s, the time interval is greater than the preset duration threshold, indicating that the current detection is overtime.

Here, in order to timely feedback the detection timeout to the user, the living body detection device may generate a first prompt message of the detection timeout, and display the prompt message.

It can be understood that, after determining the detection timeout, the living body detection device can also generate a prompt message of the detection timeout, and at the same time terminate the current living body detection process. For example, after the liveness detection is triggered by clicking the H5 jump link for bank account opening attached to the SMS content, if it is determined that the detection times out, the mobile phone can jump back to the initial SMS content interface, and the user can click the H5 jump link again. Trigger the next liveness detection.

It can be seen that the living body detection device does not need to wait for the subsequent live body detection result to fail before reminding the user, but sends a prompt message to the user in a timely manner through the display interface of the device in the case of a detection timeout, which improves the detection efficiency.

In other embodiments of the present disclosure, FIG. 7 is a schematic diagram 6 of the implementation flow of the living body detection method provided by the embodiment of the present disclosure. As shown in FIG. 7 , when the time parameter of the living body detection device is less than or equal to the preset duration threshold , after reading the k historical images to be inspected corresponding to the (i-k)th to (i-1)th images to be inspected, that is, after S103b1, the method further includes the following steps:

S103b4. If the k historical images to be inspected are all the same as the i-th image to be inspected, generate a second prompt message for detection retry.

In an implementation of the embodiment of the present disclosure, if the living body detection device uses the similarity comparison between the i-th image to be inspected and the k historical images to be inspected to realize the repeated judgment of the image to be inspected, the comparison result is the i-th image to be inspected. When the image to be inspected is the same as the k historical images to be inspected, it indicates that there is no dynamic change in the detected object, and it cannot be a legitimate living body.

For example, in the case where the detection does not time out, assuming that k is equal to 5, during the repeated judgment of the image to be inspected, if the living body detection device determines that the current image to be inspected is the same as the previous five frames of images to be inspected, it indicates that the current image to be inspected belongs to Repeated image, does not meet the initial detection conditions.

Here, in order to timely feedback to the user that there is no detection of a legitimate living body, the living body detection device may generate a second prompt message for detection retry, and display the prompt message.

For example, in the identity verification when opening a bank account through a mobile phone, if the camera is not aimed at a face, but at a static object or a static picture when collecting image data, the multiple images to be checked are repeated. If it is determined that the initial detection conditions are not met, a prompt message for detection retry can be generated, and the user can adjust the detection object in time to place the real face within the range captured by the camera.

It can be seen that the living body detection device does not need to wait for the subsequent living body detection result to fail before reminding the user, but sends a prompt message to the user through the device display interface in a timely manner when it is determined that the image is repeated, which improves the detection efficiency.

S104. If it is determined that the initial detection conditions are met, perform face detection processing on the i-th image to be checked through a preset face detection network to obtain a face detection result.

In the embodiment of the present disclosure, in the case where the live detection device determines that the detection has not timed out and the image to be inspected is not repeated, in order to avoid the problem that the detection object is not a human face when the image to be inspected is directly detected, the living body detection The device can continue to perform face detection processing on the i-th image to be checked.

In some embodiments, the liveness detection device may integrate a face detection network. The i-th image to be inspected can be input into the face detection network to determine whether there is a person in the image to be inspected based on some characteristics of the face (such as the position of the face, the pose of the face, and the size of the face, etc.) through the algorithm network. The face is judged, and then the face detection result is obtained.

In the embodiment of the present disclosure, when the living body detection device obtains the face detection result through the face detection network, there are two cases, one of which is when there is a face feature, that is, the detection object in the image to be inspected is a human face; When there is no face feature, that is, the detection object in the image to be checked is not a face.

S105. If the face detection result is that there is a face feature, perform quality detection processing according to the image parameters corresponding to the i-th image to be checked to obtain a quality detection value.

In an implementation of the embodiments of the present disclosure, after determining that the detection object is a human face, the living body detection device may analyze the image parameters of the i-th image to be inspected to determine whether the image to be inspected meets the quality requirements.

In some embodiments, the image parameter refers to a parameter pre-defined by the living body detection device and used to characterize the image quality, including at least one of light, occlusion, sharpness, and angle of the image.

Wherein, when the detection object is a face, the light of the image can be the light of the environment where the face is located, the occlusion degree can be the occlusion degree of the face, the clarity can be the clarity of the face, and the angle can be the face relative to the camera collection direction deflection angle.

Here, the living body detection device can perform quality detection processing on the light of the environment where the face is located in the image to be inspected, the clarity of the face, the degree of occlusion, and the deflection angle, and then obtain the quality scoring results of each parameter, and perform weighted calculation on each scoring result. And operation processing, the quality detection value of the image to be inspected can be obtained.

S106. If the quality detection value is greater than or equal to the preset quality threshold, determine that the i-th image to be inspected satisfies the preset quality condition.

In some embodiments of the present disclosure, the preset quality threshold may be the quality threshold of the whole face that meets the quality condition pre-defined by the living body detection device, where the quality threshold may be the light threshold of the environment where the face is located, the degree of occlusion of the face Threshold after weighted summation of threshold, face sharpness threshold, and face deflection angle threshold.

Here, the living body detection device can compare the light of the environment where the face is located with the preset light threshold, and obtain a score representing the light according to the difference between the two; compare the occlusion degree of the face with the preset occlusion degree threshold , obtain a score representing the degree of occlusion according to the degree of difference between the two; compare the sharpness of the face with the preset sharpness threshold, and obtain a score representing the sharpness according to the degree of difference between the two; The deflection angle is compared with the preset deflection angle threshold, and a score representing the degree of occlusion is obtained according to the difference between the two. The weighted sum operation is performed on the above scores to obtain the quality detection value of the i-th image to be inspected.

In some embodiments, the living body detection device compares the quality detection value obtained after the weighted summation with a preset quality threshold, and when it is determined that the quality detection value obtained by the weighted summation is greater than or equal to the preset quality threshold, the first i The image to be inspected meets the preset quality conditions.

It can be seen that, in the embodiment of the present disclosure, before the subsequent living body detection is performed, the quality judgment processing is performed on the image, so as to perform the living body detection on the to-be-detected image that meets the preset quality condition, which effectively ensures the quality of the data in the detection process and improves the performance of the detection process. detection efficiency.

In other implementations of the embodiments of the present disclosure, FIG. 8 is a schematic diagram 7 of the implementation flow of the living body detection method provided by the embodiments of the present disclosure. As shown in FIG. The image to be inspected is subjected to face detection processing, and after the face detection result is obtained, that is, after S104, the method further includes the following steps:

S107, if the face detection result is that there is no face feature, then read the n historical face detection results corresponding to the (i-n)th to (i-1)th images to be checked; wherein, n is greater than 1 and less than i the integer;

S108. If the n historical face detection results are all without face features, generate a third prompt message that the face detection fails.

In the embodiment of the present disclosure, since the face may shake out of the camera during the detection process, in order to ensure the accuracy of the face detection, the living body detection device determines that the face detection result of the i-th image to be checked is that there is no face Features, that is, when the current detection object is not a human face, the living body detection device can read the face detection results of the historical multi-frame images to be inspected, such as the n corresponding to the (i-n)th to (i-1)th images to be inspected. The historical face detection results, and then the accuracy of the face detection results is determined according to the n historical face detection results.

Here, if the n historical face detection results also all have no face features, it means that the current detection object is not a face, then in order to timely feedback the undetected face to the user, the living body detection device can generate a human face. The third prompt message of the face detection failure is displayed, and the prompt message is displayed.

For example, assuming that n is equal to 10, if the living body detection device determines that there are no facial features in the current image to be inspected and the previous 10 frames of images to be inspected, then the detected object may not be a human face, and it will prompt that no face has been detected. .

It can be seen that the living body detection device does not need to wait for the subsequent living body detection result to fail before reminding the user, but sends a prompt message to the user through the display interface of the device in a timely manner when the face detection fails, which improves the detection efficiency.

In other embodiments of the present disclosure, FIG. 9 is a schematic diagram 8 of the implementation flow of the living body detection method provided by the embodiment of the present disclosure. As shown in FIG. 9 , the living body detection device is performing quality inspection according to the image parameters corresponding to the ith image to be checked In the detection process, after the quality detection value is obtained, that is, after S105, the method further includes the following steps:

S109. If the quality detection value is less than the preset quality threshold, read the m historical quality detection values corresponding to the (i-m)th to (i-1)th images to be checked in the history; wherein, m is an integer greater than 1 and less than i ;

S110. If the m historical quality detection values are all smaller than the preset quality threshold, generate a fourth prompt message that the quality detection fails.

In this embodiment of the present disclosure, in the case where the living body detection device determines that the quality detection value of the i-th image to be inspected is less than the quality threshold, that is, the i-th image to be inspected does not meet the preset quality conditions, in order to improve the accuracy of quality detection, the living body The detection device can read the quality detection values of the historical multi-frame images to be inspected, such as the m historical quality detection values corresponding to the (i-m)th to (i-1)th images to be inspected, and then combine the m historical quality detection values to determine The accuracy of the quality judgment results.

Here, if the m historical quality detection values are all smaller than the preset quality threshold, it means that none of the image data collected by the current camera meets the quality conditions. At this time, the living body detection device generates a fourth prompt message of quality detection failure to remind the user to timely Adjust the current shooting posture.

For example, assuming that m is equal to 5, if the living body detection device determines that the quality detection value of the current image to be inspected and the quality inspection value of the previous 5 frames of images to be inspected are both less than the threshold, indicating that the data quality of the image to be inspected is poor, then the quality inspection is prompted. fail.

For example, when the user receives a notification message that the quality detection failed to be generated by the living body detection device, he can adjust the current shooting posture, including adjusting the shooting light; adjust the degree of face occlusion, such as removing obstacles such as eyes; For example, bring the face close to the camera; adjust the shooting angle, such as facing the camera, so that the living body detection device can collect images for subsequent living body detection after adjusting the posture.

In some embodiments, the above-mentioned letters k, n, and m may be set to the same numerical value, or may be set to different numerical values respectively.

It can be seen that the living body detection device does not need to wait for the subsequent living body detection result to fail before reminding the user, but sends a prompt message to the user through the display interface of the device in a timely manner when the quality detection fails, which improves the detection efficiency.

Based on the above embodiments, in an embodiment of the present disclosure, FIG. 10 is a schematic diagram 9 of the implementation flow of the living body detection method proposed in the embodiment of the present disclosure. As shown in FIG. Detection processing, the method for obtaining target detection results includes the following steps:

S102a. Determine the current living body detection result corresponding to the i-th image to be checked based on a preset living body detection network.

In the embodiment of the present disclosure, the living body detection device may determine the living body detection result corresponding to the i-th image to be checked by using the integrated living body detection network.

Here, the living body detection device may input the i-th image to be inspected into the living body detection network, so as to analyze the image to be inspected based on the characteristics of the legitimate living body through the algorithm network to obtain a probability value for characterizing the living body.

In some embodiments, the living body detection device may predefine a probability threshold, and the preset probability threshold may be used to characterize the probability threshold that the current detection object is a legitimate living body.

Here, the living body detection device may compare the living body probability value obtained through the living body detection network with the preset legal living body probability threshold value, and then determine the living body detection result of the i-th image to be checked according to the comparison result.

Wherein, when the probability value of the living body is greater than or equal to the probability threshold of a legal living body, the living body detection device can determine that the current detection object is a legal living body, that is, the living body detection result is the existence of a living body; when the living body probability value is less than the probability of a legal living body In the case of the threshold, the living body detection device determines that the current detection object is an illegal living body, which may be a face mask, or a video recorded and played in advance, that is, the living body detection result is that there is no living body.

In other embodiments, the preset probability threshold may be used to represent the probability threshold that the currently detected object is an attack object.

Here, the living body detection device may compare the living body probability value obtained through the living body detection network with the preset probability threshold of the attacking object, and then determine the living body detection result of the i-th image to be checked according to the comparison result.

Here, in the case that the living body probability value is less than the preset probability threshold of the attacking object, the living body detection device may determine that the living body detection result is the existence of a living body, that is, the current detection object is a legitimate living body; when the living body probability value is greater than or equal to the preset In the case of the probability threshold of the attacking object, the living body detection device determines that the current detection object does not exist a living body, that is, the detection object is an illegal living body, which may be a face mask, or a video recorded and played in advance, that is, the living body detection result. for the absence of living organisms.

S102b, if the current living body detection result is that there is a living body, perform a living body counting process to obtain a counting result.

S102c, if the count result is greater than or equal to the preset number threshold, determine that the target detection result is verified as passed.

In some embodiments, the living body detection device is pre-configured with a counting function module, such as a counter, for counting the living body detection results corresponding to the images to be checked, and then determining the target detection result according to the counting results.

Here, the target detection result does not refer to the in vivo detection result corresponding to the i-th image to be inspected, but the target detection result obtained by accumulating the in vivo detection results of multiple frames of images to be inspected.

Wherein, the preset result threshold is the accumulated value when the living body exists in the living body detection result of the multi-frame images to be inspected that meets the living body verification pass condition pre-defined by the living body detection device.

Here, in the case where the living body detection device determines that the living body detection result of the i-th image to be inspected is the presence of a living body, the living body detection device performs accumulation processing on the cases where the living body detection results of the multiple frames of images to be inspected are the presence of a living body through a counter, and obtains the existence of a living body. The accumulated value of the detection result, the living body detection device can compare the accumulated value with the preset result threshold, and then determine the target result according to the comparison result.

In some embodiments, if the detection result is that the accumulated value of living bodies is greater than or equal to the preset number threshold, it indicates that the living body verification pass condition is satisfied, and the living body detection device may determine that the target detection result is the verification pass.

For example, the preset number threshold is 5, and in the case where the living body detection result of the current i-th image to be inspected is that there is a living body, the updated detection result is the cumulative value of the number of living bodies. If the updated cumulative value is greater than or equal to 5, then Indicates that the living body verification pass conditions are met, and the target detection result at this time is the living body verification pass.

In some embodiments, after the living body detection device determines that the target detection result is verified, the method further includes:

S102d, generating a fifth prompt message that the living body verification is passed.

Here, the living body detection device can timely feed back the status of the living body verification to the user, wherein a prompt message of the living body verification passing can be generated and displayed to the user.

In some embodiments of the present disclosure, FIG. 11 is a schematic diagram 11 of the implementation flow of the living body detection method proposed by the embodiment of the present disclosure. As shown in FIG. 11 , the living body detection device performs the existence of living body counting processing, and after obtaining the counting result, that is, S102b After that, the method also includes the following steps:

S102e. If the count result is less than a preset number threshold, determine that the target detection result is that the verification fails.

In some embodiments, if the detection result is that the accumulated value of living bodies is less than the preset number threshold, it indicates that the living body verification failure condition is satisfied, and the living body detection device may determine that the target detection result is the verification failure.

For example, the preset number threshold is 5. In the case where the current i-th image to be inspected has a living body detection result, the updated detection result is the cumulative value of the number of living bodies. If the updated cumulative value is less than 5, it indicates that there is no living body. The living body verification pass condition is met, and the target detection result is that the living body verification fails.

In some embodiments, after the living body detection device determines that the target detection result is that the verification fails, the method further includes;

S102f. Generate a second prompt message for detection retry.

Here, the living body detection device can timely feed back to the user that the living body verification fails, wherein a prompt message for detection retry can be generated and displayed to the user.

It can be seen that, in the embodiment of the present disclosure, the living body detection device can feed back a prompt message to the user after obtaining the living body detection result based on the multi-frame to-be-detected images, and does not need to wait for the completion of the living body video detection before feeding back the detection result, which improves the detection performance. efficiency.

Based on the above-mentioned embodiment, in an embodiment of the present disclosure, FIG. 12 is a schematic diagram of an application scenario of the living body detection method proposed by the embodiment of the present disclosure. As shown in FIG. 12 , it is an exemplary application when the living body detection device is a terminal, The terminal includes a front-end module and a back-end module; among them, the front-end module is mainly responsible for calling the camera to collect video data, and cyclically extracts image frames from the collected video data according to time intervals through the frame acquisition module, and extracts image frames through the image compression module. The image frame is preprocessed by the format of picture compression, and then the image to be inspected is obtained, and then the image to be inspected is transmitted to the back-end module. The back-end module is mainly responsible for judging the quality of the images to be inspected through the quality detection module, including detection overtime judgment, image repetition judgment, face detection, face quality judgment, etc. The module performs living body detection processing to determine whether the detection object is a legitimate living body. In addition, a series of other detection processes can also be performed on the images to be inspected that meet the quality conditions through the replay attack detection module and the Track id detection module. , to achieve high-efficiency in vivo detection.

Generally speaking, in order to ensure the accuracy of the solution, the embodiments of the present disclosure provide a series of security verification means for the execution process of the entire living body detection method. The living body detection solution proposed by the embodiments of the present disclosure mainly includes a front-end data acquisition module, a data transmission module, a back-end processing module, and a data processing module. Among them, the data acquisition module is mainly used to call the camera based on http access to collect video data in the hijacking defense mode, and use Content-Security-Policy (CSP), anti-xss attack, code obfuscation, etc. Obtain the image to be inspected from the video data. The data transmission module is mainly used to transmit the acquired image to be inspected to the back-end processing module through the protocol/communication method of https encryption, digital watermarking and parameter encryption; the back-end processing module includes the application layer and the algorithm engine layer. The application layer is mainly used to perform security verification processes such as gateway authentication, interface authentication, wipe monitoring, service isolation, log security, intrusion prevention, interface current limiting, code scanning, distributed storage, and framework security. The inspection image is transmitted to the back-end algorithm engine layer; the back-end algorithm engine layer is mainly used for Track id detection, living body detection and summary information algorithm (Message-Digest Algorithm 5, MD5) detection of the image to be inspected through the application layer; After the detection is completed, the data can be stored in accordance with the storage requirements in the manner of privatization deployment, data desensitization, data permission, and data backup, thereby effectively ensuring the safe and efficient implementation of the living body detection scheme.

Based on the above embodiments, in an embodiment of the present disclosure, FIG. 13 is a schematic diagram of the execution flow of the living body detection method proposed by the embodiment of the present disclosure. As shown in FIG. 13 , the method for performing living body detection by a living body detection device includes the following steps:

S201. Determine the i-th image to be inspected from the video stream collected by the camera; wherein, i is an integer greater than 1.

S202. Does the detection time out? If no, execute S203; if yes, jump to execute S204.

S203. Are the images repeated? If no, execute S205; if yes, jump to execute S206; wherein, k is an integer greater than 1 and less than i.

S204, prompting that the detection times out.

S205, performing face detection processing.

S206. Prompt to retry the detection.

S207. Is there a human face? If yes, execute S208; if not, jump to execute S209.

S208, performing quality detection processing to obtain a quality detection value.

S209, prompting that the face cannot be detected.

S210. Is the quality detection value greater than or equal to a preset quality threshold? If yes, execute S211; if not, jump to execute S212.

S211, performing live detection.

S212, prompting that the quality detection fails.

S213: If it is a living body, determine the cumulative value of the number of living bodies.

In a case where the living body detection result of the i-th image to be inspected is that there is a living body, the current detection result is updated to be the cumulative value of the number of living bodies.

S214. Is the accumulated value greater than the preset number threshold? If yes, go to S215; if not, go to S216.

S215 , prompting that the in vivo verification is passed.

S216 , prompting that the in vivo verification fails.

It can be seen that, based on the above S201 to S216, the living body detection device performs detection based on continuous images in the living body detection process. Compared with video living body detection, the network bandwidth is occupied less, and timely feedback of the detection results can be realized. The quality inspection process is added before the inspection, which effectively ensures the data quality in the inspection process. It can be seen that the living body detection solution proposed by the embodiments of the present disclosure effectively solves the problems of large network bandwidth occupation and long detection period, and realizes high-efficiency living body detection.

Based on the above-mentioned embodiment, in an embodiment of the present disclosure, FIG. 14 is a schematic diagram of the composition and structure of a living body detection device proposed in an embodiment of the present disclosure. As shown in FIG. 14 , the living body detection device 10 includes an acquisition module 11 , a first a processing module 12, a determination module 13, a judgment module 14, a second processing module 15, a third processing module 16, a generating module 17 and a reading module 18,

The acquisition module 11 is configured to acquire the i-th image to be inspected from the video data collected by the image sensor; wherein, i is an integer greater than 1;

The first processing module 12 is configured to, in response to the i-th image to be inspected meeting a preset quality condition, perform in vivo detection processing on the i-th image to be inspected to obtain a target detection result.

In some embodiments, the acquiring module 11 is configured to extract an initial image from the video data collected by the image sensor; and perform format preprocessing on the initial image according to preset configuration parameters to obtain a processed image; and The processed image is determined as the i-th image to be inspected.

In some embodiments, the determining module 13 is configured to determine a time parameter corresponding to the i-th image to be inspected after acquiring the i-th image to be inspected from the video data collected from the image sensor.

In some embodiments, the judging module 14 is configured to judge, according to the time parameter, whether the i-th image to be inspected satisfies the initial detection condition.

In some embodiments, the second processing module 15 is configured to, if it is determined that the initial detection conditions are met, perform face detection processing on the i-th image to be checked through a preset face detection network to obtain a face Test results.

In some embodiments, the third processing module 16 is configured to perform quality detection processing according to the image parameters corresponding to the i-th image to be inspected, if the face detection result is that there is a face feature, to obtain a quality inspection value.

In some embodiments, the determining module 13 is further configured to determine that the i-th image to be inspected satisfies the preset quality condition if the quality detection value is greater than or equal to a preset quality threshold.

In some embodiments, the determining module 13 is configured to determine a time interval between the first image to be inspected and the i-th image to be inspected as the time parameter.

In some embodiments, the judging module 14 is configured to read the k historical pending images corresponding to the (i-k)th to (i-1)th images to be inspected if the time parameter is less than or equal to a preset duration threshold. wherein, k is an integer greater than 1 and less than i; and if at least one of the k historical images to be inspected is different from the i-th image to be inspected, it is determined that the initial detection is satisfied condition.

In some embodiments, the generating module 17 is configured to, after determining the time interval from the first image to be inspected to the i-th image to be inspected as the time parameter, if the time parameter is greater than the If the preset duration threshold is set, a first prompt message for detecting timeout is generated.

In some embodiments, the generating module 17 is further configured to, after reading the k historical images to be inspected corresponding to the (i-k)th to (i-1)th images to be inspected, if the k historical images to be inspected If the images are all the same as the i-th image to be inspected, a second prompt message for retrying the inspection is generated.

In some embodiments, the reading module 18 is configured to perform face detection processing on the i-th image to be inspected through a preset face detection network to obtain a face detection result, if the face detection The result is that there is no face feature, then read the n historical face detection results corresponding to the (i-n)th to (i-1)th images to be checked; wherein, n is an integer greater than 1 and less than i.

In some embodiments, the generating module 17 is further configured to generate a third prompt message that the face detection fails if the n historical face detection results are all without face features.

In some embodiments, the reading module 18 is further configured to perform quality detection processing according to the image parameters corresponding to the i-th image to be inspected to obtain a quality detection value, if the quality detection value is less than the predetermined value If the quality threshold is set, m historical quality detection values corresponding to the (i-m)th to (i-1)th historical images to be inspected are read; where m is an integer greater than 1 and less than i.

In some embodiments, the generating module 17 is further configured to generate a fourth prompt message that the quality detection fails if the m historical quality detection values are all smaller than the preset quality threshold.

In some embodiments, the first processing module 12 is configured to determine a current living body detection result corresponding to the i-th image to be checked based on a preset living body detection network; and if the current living body detection result is that there is a living body, then Counting of living bodies is performed to obtain a count result; if the count result is greater than or equal to a preset number threshold, it is determined that the target detection result is verified.

In some embodiments, the first processing module 12 is further configured to input the i-th image to be inspected into the preset living body detection network to obtain a living body probability value; and if the living body probability value is greater than or equal to a predetermined value If a probability threshold is set, it is determined that the current living body detection result is the presence of a living body; and if the living body probability value is less than the preset probability threshold, it is determined that the current living body detection result is that there is no living body.

In some embodiments, the first processing module 12 is further configured to, after performing the live body counting process and obtaining the counting result, if the counting result is less than a preset number threshold, determine that the target detection result is not verified. pass.

In some embodiments, the generating module 17 is further configured to, after performing a living body detection process on the i-th image to be inspected and obtaining a target detection result, in response to the target detection result being that the verification is passed, generate a successful living body detection and in response to the target detection result being that the verification fails, generating a second prompt message for detection retry.

In some embodiments, the image parameters include at least one of light, occlusion, sharpness, and angle.

In the embodiment of the present disclosure, further, FIG. 15 is a schematic diagram of the composition structure of the living body detection device proposed by the embodiment of the present disclosure. As shown in FIG. 15 , the living body detection device 20 proposed by the embodiment of the present disclosure may further include a processor 21 , a memory 22 storing executable instructions of the processor 21 , further, the living body detection device 20 may further include a communication interface 23 , and a bus 24 for connecting the processor 21 , the memory 22 and the communication interface 23 .

In the embodiment of the present disclosure, the above-mentioned processor 21 may be an application specific integrated circuit (ASIC), a digital signal processor (Digital Signal Processor, DSP), a digital signal processing device (Digital Signal Processing Device, DSPD) ), Programmable Logic Device (ProgRAMmable Logic Device, PLD), Field Programmable Gate Array (Field Prog RAMmable Gate Array, FPGA), Central Processing Unit (Central Processing Unit, CPU), controller, microcontroller, microprocessor at least one of them. It can be understood that, for different devices, the electronic device used to implement the function of the processor may also be other, which is not specifically limited in the embodiment of the present disclosure. The living body detection device 20 may also include a memory 22, which may be connected to the processor 21, wherein the memory 22 is used to store executable program codes, which include computer operating instructions, and the memory 22 may include high-speed RAM memory, or may Also included is non-volatile memory, eg, at least two disk drives.

In the embodiment of the present disclosure, the bus 24 is used to connect the communication interface 23 , the processor 21 and the memory 22 and the mutual communication among these devices.

In the embodiment of the present disclosure, the memory 22 is used to store instructions and data.

Further, in the embodiment of the present disclosure, the above-mentioned processor 21 is configured to obtain the i-th image to be inspected from the video data collected by the image sensor; wherein, i is an integer greater than 1; in response to the i-th image to be inspected If the image satisfies the preset quality condition, the i-th image to be inspected is subjected to in vivo detection processing to obtain a target detection result.

In practical applications, the above-mentioned memory 22 may be a volatile memory (volatile memory), such as a random access memory (Random-Access Memory, RAM); or a non-volatile memory (non-volatile memory), such as a read-only memory (Read-Only Memory, ROM), flash memory (flash memory), hard disk (Hard Disk Drive, HDD) or solid-state drive (Solid-State Drive, SSD); or a combination of the above-mentioned types of memory, and send it to the processor 15 Provide instructions and data.

In addition, each functional module in this embodiment may be integrated into one, or each unit may exist physically alone, or two or more units may be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware, or can be implemented in the form of software function modules.

If the integrated unit is implemented in the form of software function modules and is not sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment is essentially or correct. Part of the contribution made by the prior art or all or part of the technical solution can be embodied in the form of a software product, the computer software product is stored in a storage medium, and includes several instructions to make a computer device (which can be a personal A computer, a server, or a network device, etc.) or a processor (processor) executes all or part of the steps of the method in this embodiment. The aforementioned storage medium includes: a U disk, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk and other mediums that can store program codes.

An embodiment of the present disclosure provides a living body detection device, which can first obtain an i-th image to be inspected from video data collected by an image sensor; The i-th image to be inspected is subjected to live detection processing to obtain a target detection result. In this way, the detection process is based on continuous images in the process of living body detection. Compared with video living body detection, the network bandwidth is less occupied, and timely feedback of the detection results can be realized, and the quality detection process is added before the subsequent living body detection, which effectively ensures the detection. data quality in the process. It can be seen that the living body detection solution proposed by the embodiments of the present disclosure effectively solves the problems of large network bandwidth occupation and long detection period, and realizes high-efficiency living body detection.

An embodiment of the present disclosure provides a computer-readable storage medium on which a program is stored, and when the program is executed by a processor, implements the above-mentioned method for detecting a living body.

Specifically, a program instruction corresponding to a living body detection method in this embodiment may be stored on a storage medium such as an optical disk, a hard disk, a U disk, etc. When the program instruction corresponding to a living body detection method in the storage medium is stored in a When the electronic device reads or is executed, it includes the following steps:

Obtain the i-th image to be inspected from the video data collected by the image sensor; wherein, i is an integer greater than 1;

In response to the i-th image to be inspected meeting a preset quality condition, a living body detection process is performed on the i-th image to be inspected to obtain a target detection result.

Correspondingly, an embodiment of the present disclosure further provides a computer program product, wherein the computer program product includes computer-executable instructions, and the computer-executable instructions are used to implement the steps in the living body detection method provided by the embodiment of the present disclosure.

As will be appreciated by one skilled in the art, embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable storage media having computer-usable program code embodied therein, including but not limited to disk storage, optical storage, and the like.

The present disclosure is described with reference to schematic flowchart illustrations and/or block diagrams of implementations of methods, apparatus (systems), and computer program products according to embodiments of the present disclosure. It will be understood that each process and/or block in the schematic flowchart illustrations and/or block diagrams, and combinations of processes and/or blocks in the schematic flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to the processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing device to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing device produce Means for implementing the functions specified in a process or processes and/or a block or blocks in the block diagrams.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory result in an article of manufacture comprising instruction means, the instructions An apparatus implements the functions specified in a flow or flows of the implementation flow diagram and/or a block or blocks of the block diagram.

These computer program instructions can also be loaded on a computer or other programmable data processing device to cause a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process such that The instructions provide steps for implementing the functions specified in the flow or blocks of the implementing flow diagram and/or the block or blocks of the block diagram.

The above descriptions are merely preferred embodiments of the present disclosure, and are not intended to limit the protection scope of the present disclosure.

Industrial Applicability

In the embodiment of the present disclosure, the i-th image to be inspected is determined from the video data collected by the image sensor; in response to the i-th image to be inspected meeting a preset quality condition, a living body detection process is performed on the i-th image to be inspected to obtain a target detection result . High-efficiency in vivo detection is achieved.

Claims

A method for detecting a living body, the method comprising:

Obtain the i-th image to be inspected from the video data collected by the image sensor; wherein, i is an integer greater than 1;

In response to the i-th image to be inspected meeting a preset quality condition, a living body detection process is performed on the i-th image to be inspected to obtain a target detection result.
The method according to claim 1, wherein the acquiring the i-th image to be inspected from the video data collected by the image sensor comprises:

extracting an initial image from the video data collected by the image sensor;

Perform format preprocessing on the initial image according to preset configuration parameters to obtain a processed image;

The processed image is determined as the i-th image to be inspected.
The method according to claim 1, wherein after acquiring the i-th image to be inspected from the video data collected by the image sensor, the method further comprises:

determining the time parameter corresponding to the i-th image to be inspected, and judging whether the i-th image to be inspected satisfies the initial detection condition according to the time parameter;

If it is determined that the initial detection conditions are met, performing face detection processing on the i-th image to be inspected through a preset face detection network to obtain a face detection result;

If the face detection result is that there is a face feature, then perform quality detection processing according to the image parameters corresponding to the i-th image to be checked to obtain a quality detection value;

If the quality detection value is greater than or equal to a preset quality threshold, it is determined that the i-th image to be inspected satisfies the preset quality condition.
The method according to claim 3, wherein the determining the time parameter corresponding to the i-th image to be inspected comprises:

Determining the time interval from the first image to be inspected to the i-th image to be inspected as the time parameter;

Correspondingly, judging whether the i-th image to be inspected satisfies the initial detection condition according to the time parameter includes:

If the time parameter is less than or equal to the preset duration threshold, read the k historical images to be inspected corresponding to the (i-k)th to (i-1)th images to be inspected; wherein, k is an integer greater than 1 and less than i ;

If at least one of the k historical images to be inspected is different from the i-th image to be inspected, it is determined that the initial detection condition is satisfied.
The method according to claim 4, wherein after the time interval between the first image to be inspected and the image to be inspected i is determined as the time parameter, the method further comprises:

If the time parameter is greater than the preset duration threshold, a first prompt message for detecting a timeout is generated.
The method according to claim 4, wherein after reading the k historical images to be inspected corresponding to the (i-k)th to (i-1)th images to be inspected, the method further comprises:

If the k historical images to be inspected are all the same as the i-th image to be inspected, a second prompt message for detection retry is generated.
The method according to claim 3, wherein, after performing face detection processing on the i-th image to be inspected through a preset face detection network to obtain a face detection result, the method further comprises:

If the face detection result is that there is no face feature, then read the n historical face detection results corresponding to the (i-n)th to (i-1)th images to be checked; wherein, n is greater than 1 and less than i the integer;

If all the n historical face detection results do not have a face feature, a third prompt message that the face detection fails is generated.
The method according to claim 3, wherein after the quality detection process is performed according to the image parameters corresponding to the i-th image to be inspected to obtain a quality detection value, the method further comprises:

If the quality detection value is smaller than the preset quality threshold, read the m historical quality detection values corresponding to the (i-m)th to (i-1)th images to be checked in history; where m is greater than 1 and less than i the integer;

If the m historical quality detection values are all smaller than the preset quality threshold, a fourth prompt message for quality detection failure is generated.
The method according to claim 1, wherein the performing a living body detection process on the i-th image to be inspected to obtain a target detection result comprises:

determining the current living body detection result corresponding to the i-th image to be inspected based on a preset live body detection network;

If the current living body detection result is that there is a living body, perform a living body counting process to obtain a counting result;

If the count result is greater than or equal to a preset number threshold, it is determined that the target detection result is verified.
The method according to claim 9, wherein the determining the current living body detection result corresponding to the i-th image to be checked based on a preset living body detection network comprises:

Inputting the i-th image to be inspected into the preset living detection network to obtain a living probability value;

If the living body probability value is greater than or equal to a preset probability threshold, it is determined that the current living body detection result is the presence of a living body;

If the living body probability value is smaller than the preset probability threshold value, it is determined that the current living body detection result is that there is no living body.
The method according to claim 9, wherein, after the living body counting process is performed and the counting result is obtained, the method further comprises:

If the count result is less than the preset number threshold, it is determined that the target detection result is a failed verification.
The method according to claim 11, wherein, after the i-th image to be inspected is subjected to in vivo detection processing to obtain a target detection result, the method further comprises at least one of the following:

In response to the target detection result being that the verification is passed, generating a fifth prompt message indicating that the living body detection is successful;

In response to the target detection result being that the verification fails, a second prompt message for detection retry is generated.
The method of claim 3, wherein the image parameters include at least one of light, occlusion, sharpness, and angle.
A living body detection device, the living body detection device includes an acquisition module and a first processing module,

The acquisition module is configured to acquire the i-th image to be inspected from the video data collected by the image sensor; wherein, the i is an integer greater than 1;

The first processing module is configured to, in response to the i-th image to be inspected meeting a preset quality condition, perform in vivo detection processing on the i-th image to be inspected to obtain a target detection result.
The living body detection device according to claim 14, wherein,

The acquisition module is configured to extract an initial image from the video data collected by the image sensor; and perform format preprocessing on the initial image according to preset configuration parameters to obtain a processed image; and determine the processed image is the i-th image to be inspected.
The living body detection device according to claim 14, wherein the living body detection device further comprises a determination module, a judgment module, a second processing module and a third processing module,

The determining module is configured to determine the time parameter corresponding to the i-th image to be inspected after acquiring the i-th image to be inspected from the video data collected from the image sensor;

The judging module is configured to judge whether the i-th image to be inspected satisfies the initial detection condition according to the time parameter;

The second processing module is configured to, if it is determined that the initial detection conditions are met, perform face detection processing on the i-th image to be inspected through a preset face detection network to obtain a face detection result;

The third processing module is configured to perform quality detection processing according to the image parameters corresponding to the i-th image to be checked to obtain a quality detection value if the face detection result is that there is a face feature;

The determining module is further configured to determine that the i-th image to be inspected satisfies the preset quality condition if the quality detection value is greater than or equal to a preset quality threshold.
The living body detection device according to claim 16, wherein,

The determining module is configured to determine the time interval from the first image to be inspected to the i-th image to be inspected as the time parameter;

Correspondingly, the judging module is configured to read k historical images to be inspected corresponding to the (i-k)th to (i-1)th images to be inspected if the time parameter is less than or equal to the preset duration threshold; wherein , k is an integer greater than 1 and less than i; and if at least one of the k historical images to be inspected is different from the i-th image to be inspected, it is determined that the initial detection condition is satisfied.
The living body detection device according to claim 17, wherein the living body detection device further comprises a generating module,

The generating module is configured to, after determining the time interval from the first image to be inspected to the i-th image to be inspected as the time parameter, if the time parameter is greater than the preset duration threshold, generate The first prompt message for the detection timeout.
The living body detection device according to claim 17, wherein,

The generating module 17 is further configured to read the k historical images to be inspected corresponding to the (i-k)th to (i-1)th images to be inspected, if the k historical images to be inspected are the same as the i The images to be inspected are the same, and a second prompt message for retrying the inspection is generated.
The living body detection device according to claim 16, wherein the living body detection device further comprises a reading module,

The reading module is configured to perform face detection processing on the i-th image to be checked through a preset face detection network, and after obtaining a face detection result, if the face detection result is that there is no face feature , then read the n historical face detection results corresponding to the (i-n) to (i-1) images to be inspected; wherein, n is an integer greater than 1 and less than i;

The generating module is further configured to generate a third prompt message that the face detection fails if all the n historical face detection results do not have face features.
The living body detection device according to claim 16, wherein,

The reading module is further configured to perform quality detection processing according to the image parameters corresponding to the i-th image to be checked, and after obtaining a quality detection value, if the quality detection value is less than the preset quality threshold, read The m historical quality detection values corresponding to the historical (i-m)th to (i-1)th images to be inspected; wherein, m is an integer greater than 1 and less than i;

The generating module 17 is further configured to generate a fourth prompt message that the quality detection fails if the m historical quality detection values are all less than the preset quality threshold.
The living body detection device according to claim 14, wherein,

The first processing module is configured to determine a current living body detection result corresponding to the i-th image to be inspected based on a preset living body detection network; and if the current living body detection result is the existence of living bodies, perform a living body counting process to obtain Counting result; if the counting result is greater than or equal to the preset number threshold, it is determined that the target detection result is verified.
The living body detection device according to claim 22, wherein,

The first processing module is further configured to input the i-th image to be inspected into the preset living body detection network to obtain a living body probability value; and if the living body probability value is greater than or equal to a preset probability threshold, determine the The current living body detection result is that there is a living body; and if the living body probability value is less than the preset probability threshold, it is determined that the current living body detection result is that there is no living body.
The living body detection device according to claim 22, wherein,

The first processing module is further configured to, after performing the counting process of the existence of living body and obtaining the counting result, if the counting result is less than the preset number threshold, determine that the target detection result is not passed the verification.
The living body detection device according to claim 24, wherein,

The generating module is further configured to generate a fifth prompt message of successful living body detection in response to the target detection result being that the verification is passed after the i-th image to be inspected is subjected to in vivo detection processing to obtain a target detection result; and In response to the target detection result being that the verification fails, a second prompt message for detection retry is generated.
The living body detection device according to claim 16, wherein,

The image parameters include at least one of light, occlusion, sharpness, and angle.
A living body detection device, the living body detection device comprises a processor and a memory storing instructions executable by the processor, when the instructions are executed by the processor, the implementation of any one of claims 1-13 is implemented. method described.
A computer-readable storage medium having a program stored thereon and applied to a living body detection device, when the program is executed by a processor, the method according to any one of claims 1-13 is implemented.
A computer program, comprising computer-readable codes, in the case that the computer-readable codes are executed in an electronic device and executed by a processor in the electronic device, to implement the method described in any one of claims 1 to 13 method.
A computer program product which, when run on a computer, causes the computer to perform the method of any one of claims 1 to 13.