CN111738176A - A living body detection model training, living body detection method, device, equipment and medium - Google Patents

Abstract

The embodiment of the specification discloses a living body detection model training method, a living body detection device, equipment and a medium, wherein the living body detection model training method comprises the following steps: obtaining a sample set for model training, wherein samples in the sample set comprise positive samples and negative samples, and each sample comprises at least two face pictures with different distances; determining an optical flow graph corresponding to each sample; and training a living body detection model by using the optical flow graph corresponding to each sample.

Description

A living body detection model training, living body detection method, device, equipment and medium

technical field

The present application relates to the field of computer technology, and in particular, to a training of a living body detection model, a living body detection method, an apparatus, a device and a medium.

Background technique

Liveness detection is an important application of computer technology, and how to improve the effect and efficiency of liveness detection has become an important topic.

In view of this, there is a need for more effective and efficient living detection schemes.

SUMMARY OF THE INVENTION

The embodiments of the present specification provide a living body detection model training, a living body detection method, an apparatus, a device, and a medium, so as to solve the technical problem of how to perform the living body detection more effectively and efficiently.

In order to solve the above-mentioned technical problems, the embodiments of this specification adopt the following technical solutions:

The embodiments of this specification provide a method for training a live detection model, including:

Obtain a sample set for model training, the samples in the sample set include positive samples and negative samples, and each sample includes at least two face pictures with different distances;

Determine the optical flow map corresponding to each sample;

Using the optical flow map corresponding to each sample to train a living body detection model.

The embodiments of this specification provide a first method for detecting a living body, including:

Obtaining a picture group including at least two face pictures of different distances, the face pictures including the face area of the living body detection object;

determining an optical flow map corresponding to the picture group;

Input the optical flow map into the living body detection model obtained by the above training, and determine whether the living body detection object is a living body according to the output data of the living body detection model.

The embodiments of this specification provide a second method for detecting a living body, including:

collecting a picture group including at least two face pictures at different distances, the face pictures including the face region of the living body detection object;

Send the picture group to the living body detection end, so that the living body detection end determines the optical flow map corresponding to the picture group; and, input the optical flow map into the living body detection model obtained by the above training, The output data of the model determines whether the living body detection object is a living body.

The embodiments of this specification provide a live detection model training device, including:

a sample module, configured to obtain a sample set for model training, the samples in the sample set include positive samples and negative samples, and each sample includes at least two face pictures with different distances;

The optical flow module is used to determine the optical flow map corresponding to each sample;

A training module, configured to train a living body detection model by using the optical flow graph corresponding to each sample.

The embodiments of this specification provide a living body detection device, the device is configured with the living body detection model obtained by the above training, including:

A picture module, configured to obtain a picture group including at least two face pictures at different distances, the face pictures including the face region of the living body detection object;

an optical flow module, configured to determine an optical flow map corresponding to the picture group;

The detection module is configured to input the optical flow map into the living body detection model, and determine whether the living body detection object is a living body according to the output data of the living body detection model.

The embodiments of this specification provide a living body detection device, including:

A picture module, configured to collect a picture group including at least two face pictures at different distances, the face pictures including the face region of the living body detection object;

a sending module, configured to send the picture group to the living body detection end, so that the living body detection end determines the optical flow map corresponding to the picture group; and, input the optical flow map into any one of claims 1 to 7 item, and determine whether the object of living body detection is a living body according to the output data of the living body detection model.

The embodiments of this specification provide a device for training a living body detection model, including:

at least one processor;

as well as,

a memory communicatively coupled to the at least one processor;

in,

The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the above-described method for training a living body detection model.

The embodiments of this specification provide a living body detection device, including:

at least one processor;

as well as,

a memory communicatively coupled to the at least one processor;

in,

The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the above-mentioned first method of living body detection.

The embodiments of this specification provide a living body detection device, including:

at least one processor;

as well as,

a memory communicatively coupled to the at least one processor;

in,

The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute the above-mentioned second method of living body detection.

An embodiment of the present specification provides a computer-readable storage medium, where the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, implements the above-mentioned method for training a living body detection model.

An embodiment of the present specification provides a computer-readable storage medium, where the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, implements the above-mentioned first method for training a living body detection model.

An embodiment of the present specification provides a computer-readable storage medium, where the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, implements the above-mentioned second method for training a living body detection model.

The above-mentioned at least one technical solution adopted in the embodiments of this specification can achieve the following beneficial effects:

Calculate the optical flow of face pictures at different distances and generate an optical flow map. The optical flow map is used to train the living body detection model and use it for living body detection, which can reflect the dynamic changes of the face and the perspective effect, and improve the living body detection effect and efficiency.

Description of drawings

In order to more clearly illustrate the embodiments of the present specification or the technical solutions in the prior art, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments of the present specification or the prior art. Obviously, the drawings in the following description These are only some embodiments described in this specification, and for those skilled in the art, other drawings can also be obtained from these drawings without any creative effort.

FIG. 1 is a schematic diagram of an execution body of a method for training a living body detection model provided by the first embodiment of the present specification.

FIG. 2 is a schematic flowchart of a method for training a living body detection model provided by the first embodiment of the present specification.

FIG. 3 is a schematic diagram of training a living body detection model provided by the first embodiment of this specification.

FIG. 4 is a schematic flowchart of a method for detecting a living body provided in a second embodiment of the present specification.

FIG. 5 is a schematic diagram of a display in the second embodiment of the present specification.

FIG. 6 is another schematic diagram of display in the second embodiment of the present specification.

FIG. 7 is another schematic diagram of display in the second embodiment of the present specification.

FIG. 8 is a schematic flowchart of a method for detecting a living body provided by a third embodiment of the present specification.

FIG. 9 is a schematic structural diagram of an apparatus for training a living body detection model provided in a fourth embodiment of the present specification.

FIG. 10 is a schematic structural diagram of a living body detection device provided in a fifth embodiment of the present specification.

FIG. 11 is a schematic structural diagram of a living body detection device provided in a sixth embodiment of the present specification.

Detailed ways

In order to make those skilled in the art better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings in the embodiments of this specification. Obviously, the described The embodiments are only a part of the embodiments of the present application, but not all of the embodiments. Based on the embodiments of the present specification, all other embodiments obtained by persons of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

In the prior art, liveness detection has become a relatively common requirement. For example, in scenarios such as face-swiping login, face-swiping payment, etc., liveness detection is often required. Generally speaking, it is necessary to collect pictures or videos of the detection object for live detection, and there are many models or algorithms for live detection, such as single-frame silent live algorithms or action live algorithms based on deep learning models (that is, requiring live detection algorithms for actions such as blinking, shaking your head, etc.). On the other hand, there are currently some deceptive behaviors for liveness detection, such as liveness detection through software-generated videos or pictures, or reproduced pictures or printed pictures. For example, for the above-mentioned single-frame silent living algorithm, although it can effectively intercept screens and retake photos with obvious borders and material characteristics, the single-frame algorithm has poor recognition effect for high-definition screen photos; The living algorithm can also recognize high-definition screen photos or remade or printed photos to a certain extent, but it is easily broken by videos generated by face generation software that contain actions (such as blinking, shaking, opening, etc.).

The first embodiment of this specification provides a method for training a living body detection model. The execution subject of this embodiment may be a terminal (including but not limited to a mobile phone, a computer, a pad, etc.) or a server or a corresponding living body detection model training platform or system or The operating system, etc., that is, the execution body, can be various, and the execution body can be set, used or transformed as required. In addition, there may also be a third-party application program to assist the executive body to execute this embodiment. For example, as shown in FIG. 1 , the method for training a living body detection model in this embodiment may be executed by a server, and an application program (corresponding to the server) may be installed on a terminal (held by the user), the terminal or the application Data transmission can be performed between the program and the server, data collection, input or output, or page or information display (to the user) can be performed through a terminal or an application program, thereby assisting the server to execute the method for training a living body detection model in this embodiment.

As shown in FIG. 2 and FIG. 3 , the method for training a living body detection model provided by this embodiment includes:

S101: (execution subject) obtain a sample set for model training, the samples in the sample set include positive samples and negative samples, and each sample includes at least two face pictures with different distances;

In this embodiment, a sample set can be obtained, the sample set includes several samples, the samples can include positive samples and negative samples, and each sample includes at least two face pictures with different distances (the face pictures can be from a corresponding database) . Further, each face image should contain the complete face area. In this embodiment, the positive sample is a sample from a living body, that is, the face picture in the positive sample is from a living body (the living body in this specification generally refers to a natural person), for example, the row of “living body” in FIG. 3 . Negative samples are samples that are not from a living body, that is, the face pictures in the negative samples are not from a living body, including but not limited to the face pictures in the negative samples are generated by software or reproduced from screens or photos, such as in Figure 3 Two lines for "Photo" and "Screen".

In this embodiment, the number of face pictures included in each sample is not specifically limited, and may be two or more than two. Generally speaking, the face pictures contained in each positive sample should be from the same living body, and the face pictures contained in each negative sample should also be the face of the same person, even if it is generated by software or retaken.

In this embodiment, whether the distance between any two facial pictures is the same or different may be determined according to the ratio of the region of interest in each facial picture to the facial picture. Specifically, the face region in the face picture (the face region can be selected by an existing face detection algorithm, and the shape of the selected face region includes but not limited to a rectangle) can be used as the region of interest, and each calculation The ratio of the region of interest in the face picture to the face picture, including but not limited to calculating the ratio of the area of the region of interest in each face picture to the area of the face picture, or calculating the ratio of each face picture The ratio of the width of the region of interest in the picture to the width of the face picture, or the ratio of the length of the region of interest in each face picture to the length of the face picture is calculated. The above-mentioned units of area/width/length include, but are not limited to, centimeters, pixels, inches, and the like. The above ratio can be calculated using (existing) face detection algorithms or models, including but not limited to the MTCNN algorithm.

The distance between different face pictures can be determined according to the above ratio. Specifically, if the above ratio is smaller, the larger the distance is determined. For example, the above ratio in picture A is 0.5, and the above ratio in picture B is 0.3, then it is determined that the distance of picture A is smaller than the distance of picture B. It should be noted that when comparing the above ratios of any two face pictures, the ratios with the same meaning should be compared, for example, the two pictures both adopt the area ratio, or both adopt the width ratio, or both adopt the length ratio, etc.

"Different distances" in this embodiment are equivalent to different shooting distances. For positive samples, face pictures at different distances may be obtained due to the difference between the living body and the shooting device. For negative samples, due to the above ratio Differently, the face pictures in the negative samples at least "look" at different shooting distances, or represent different shooting distances.

S103: (execution subject) determine the optical flow map corresponding to each sample;

In this embodiment, for each sample, its corresponding optical flow map can be determined. Specifically, determining the optical flow map corresponding to each sample includes: for any sample, aligning each face picture included in the sample, and using each face picture aligned by the sample to determine the optical flow map corresponding to the sample.

In this embodiment, for any sample, aligning the facial pictures included in the sample includes:

S1031: For any sample, perform the first-stage alignment of each facial image included in the sample according to the facial key points in each facial image included in the sample;

Specifically, for any sample, extract the key points of each face picture in the sample, and perform the first-stage alignment of each face picture in the sample according to the key points (the first-stage alignment may also be referred to as "coarse alignment" ). Extracting keypoints can use existing algorithms or models.

S1033: According to the machine vision feature of each facial image included in the sample, perform a second-stage alignment on each facial image after the first alignment of the sample.

After the face pictures of the sample are aligned in the first stage, the machine vision features of the face pictures in the sample can be extracted, and the face pictures in the sample can be aligned in the second stage according to the machine vision features (the second stage). Stage alignment may also be referred to as "fine alignment"). The machine vision features may include at least one of SIFT features, HOG features, SURF features, ORB features, LBP features, and HAAR features.

For any sample, after aligning the face pictures in the sample, the optical flow of the (aligned) face pictures in the sample can be calculated, according to the optical flow of the (aligned) face pictures in the sample An optical flow graph corresponding to each face picture in the sample is generated, that is, an optical flow graph corresponding to the sample.

Using the above content, the optical flow map corresponding to each sample (including positive samples and negative samples) can also be determined.

S105: (Execution subject) Use the optical flow graph corresponding to each sample to train a living body detection model.

In this embodiment, after the optical flow map corresponding to each sample is generated, the living body detection model can be trained by using the optical flow map corresponding to each sample. Wherein, using the optical flow graph corresponding to each sample to train the living body detection model includes: using the optical flow graph corresponding to each sample to train a classification model (or optical flow graph classification model) to obtain a living body detection model.

The above classification models include but are not limited to CNN classification models, PCA models, etc., which are not specifically limited in this embodiment.

In this embodiment, each sample contains face pictures at different distances. If the face pictures in the sample are from a living body, different perspective effects will be formed due to the different distances between the living body and the photographing device. The flow graph can accurately represent the perspective effect, so in this embodiment, the optical flow graph is used to train the living body detection model, so that the obtained living body detection model can perform living body detection based on the above-mentioned perspective effect. Including, but not limited to, non-living videos or pictures such as videos or pictures generated by software, re-taken pictures or printed pictures, etc., cannot show the above perspective effect, so the living body detection model obtained in this embodiment can accurately detect whether there is a living body , the live detection effect is good and the efficiency is high. In this embodiment, the facial image can be aligned in two stages to generate an optical flow map, so that the generated optical flow map (for the above perspective effect) has a better performance effect, and improves the living body detection effect of the obtained living body detection model. efficiency.

The second embodiment of the present specification provides a living body detection method, which can be regarded as an application of the living body detection model obtained in the first embodiment. The execution body of this embodiment may be a terminal (including but not limited to a mobile phone, a computer, a pad, etc.) or a server or a corresponding living body detection platform or system or operating system, etc., that is, the execution body may be various, and can be set as required , use or transform the execution body. In addition, there may also be a third-party application program to assist the executive body to execute this embodiment. For example, as shown in FIG. 1 , the living body detection method in this embodiment may be executed by a server, and an application program (corresponding to the server) may be installed on a terminal (held by the user), and the terminal or application program is associated with Data transmission can be performed between servers, data collection, input or output, or page or information display (to the user) can be performed through a terminal or an application program, thereby assisting the server to execute the living body detection method in this embodiment.

As shown in FIG. 4 , the live detection method provided in this embodiment includes:

S201: (execution subject) acquiring a picture group including at least two face pictures with different distances, the face pictures including the face region of the living body detection object;

In this embodiment, a picture group including at least two face pictures at different distances (equivalent to different shooting distances) may be acquired. The method described in 1.1 or 1.2 or 1.3 below can be used to obtain a picture group including at least two face pictures with different distances (this embodiment is not limited to the method described in 1.1 or 1.2 or 1.3):

1.1. Video decomposition

In this embodiment, a shooting device (which may include but is not limited to a camera, the same below) can shoot a video or picture for a living body detection object (according to the different living body detection results, the living body detection object here may be a living body or a non-living body) , referred to as "non-living body", such as remake or generated video or picture, the same below). Specifically, the photographing device may start photographing after detecting the living body detection object. The photographing device may send the photographed video to the execution body (“execution body” for short) in this embodiment.

After receiving the above-mentioned shooting video, the execution subject can decompose the shooting video to obtain several frames of images, and select at least two face pictures with different distances from the several frames of images to form a picture group. Whether the distances of each frame of pictures are different can be determined by using the method provided in the first embodiment, the same below.

1.2. Receive pictures

In this embodiment, the photographing device may send the photographed facial image to the execution subject every time a facial image is photographed for the living body detection object. Each time the execution subject receives a face picture, it sends an instruction to the shooting device, so that the shooting device issues prompt information, and the prompt information is used to prompt the living body detection object to approach or stay away from the shooting device. For example, the prompt information can be It's "please move closer to or away from the camera". The photographing device may issue the prompt information in the form of text or audio, the same below. For example, the photographing device has a display screen, and the prompt information is published in text form through the display screen, as shown in FIG. 5; and/or the photographing device has an audio module, and the prompt information is published in audio form through the audio module.

It can be expected that if the living body detection object is a living body, after receiving (for example, seeing or hearing, the same below) the prompt information, it will approach or move away from the photographing device; if the living body detection object is a non-living body, it should be If the holder of the object is placed in front of the photographing device for photographing, if the holder of the living body detection object receives the prompt information, the owner of the living body detection object will also move the living body detection object closer to or away from the photographing device.

After the photographing device releases the prompt information, it will photograph the next face picture and send it to the execution subject.

The photographing device repeats the above process, that is, "take and send the last face picture-receive instruction and release prompt information-take and send the next face picture" ("previous" and "next" are adjacent two. Zhang, the same below), so that the execution subject can receive a series of face pictures, and select at least two face pictures with different distances from them to form a picture group.

1.3. Receive pictures and judge the distance

In this embodiment, the photographing device may send the photographed facial image to the execution subject every time a facial image is photographed for the living body detection object. Each time the execution subject receives a face picture, it determines the prompt information according to the ratio of the region of interest in the face picture to the picture (same as the first embodiment), and sends the prompt information to the shooting device to The photographing device is caused to issue the prompt information, where the prompt information is used to prompt the living body detection object to approach or move away from the photographing device.

Specifically, a threshold can be set, and each time the execution subject receives a face picture, it determines the ratio of the region of interest in the face picture to the picture. If the ratio is greater than (or greater than or equal to) the threshold, it means that the living body The detection object is closer to the shooting device, and the prompt information can be used to prompt the living body detection object to stay away from the shooting device. For example, the prompt information can be "Please stay away from the camera", as shown in Figure 6; if the ratio is less than or equal to (or less than or equal to) ) the threshold value, indicating that the living body detection object is far away from the shooting device, and the prompt information can be used to prompt the living body detection object to approach the shooting device, for example, the prompt information can be “please approach the camera”, as shown in FIG. 7 .

After the photographing device releases the prompt information, it will photograph the next face picture and send it to the execution subject.

The shooting device repeats the above process, that is, "shooting and sending the last face picture - receiving instruction and issuing prompt information - shooting and sending the next face picture", so that the execution subject can receive a series of face pictures, and select at least one of them. Two pictures of faces at different distances to form a picture group.

In particular, in 1.2 or 1.3, after the photographing device issues the prompt information, it can detect whether the living body detection object has moved, and only after the movement has occurred, the next face picture can be taken. If the living body detection object does not move after the prompt information is issued, the prompt information is continuously issued to prevent the non-living body holder from moving the living body detection object and evading the living body detection when the living body detection object is a non-living body.

In 1.2, the prompt information does not limit the living body detection object to be close to or away from the shooting device. In 1.3 (by proportion), the distance between the living body detection object and the shooting device is judged, and then the prompt information limits the living body detection object to approach or stay away from the shooting device, so that the facial image The face area in the image is not too large or too small, making the face image better. In 1.3, if the living body detection object is close to the shooting device, it can be prompted to stay away from the shooting device, otherwise, it can be prompted to approach the shooting device, that is, it can adapt to the distance between the living body detection object and the shooting device, and improve the quality of face pictures.

In 1.2 and 1.3, if the execution subject has received a sufficient number of face images that satisfy different distances, it may instruct the image capture to no longer capture face images.

The face images in the above-mentioned group of pictures should satisfy the (complete) face region containing the living body detection object, even if the living body detection object is non-living body, it can be foreseen that it also contains the face region for use in the living body detection.

In this embodiment, it can be determined whether the distance between any two facial images is the same or different according to the ratio of the region of interest in each facial image to the facial image. Refer to the content in the first embodiment. .

S203: (execution subject) determine the optical flow map corresponding to the picture group;

After the picture group is obtained, the optical flow map corresponding to each face picture in the picture group can be determined, and the optical flow map corresponding to the picture group is also determined, that is, each of the face pictures is aligned, and each of the aligned faces is used. The partial picture determines the optical flow map corresponding to the picture group. The first-stage alignment is performed on each of the facial pictures according to the facial key points in each of the facial pictures; according to the machine vision features of each of the facial pictures, each The face image is aligned in the second stage. The machine vision features include at least one of SIFT features, HOG features, SURF features, ORB features, LBP features, and HAAR features.

Using the aligned face pictures to determine the optical flow map corresponding to the picture group includes: calculating the optical flow of each of the face pictures; generating an optical flow corresponding to the picture group according to the optical flow of each of the face pictures flow graph.

For specific content, refer to the content in the first embodiment.

S205: (Execution subject) Input the optical flow graph into the living body detection model obtained in the first embodiment, and determine whether the living body detection object is a living body according to the output data of the living body detection model.

After the optical flow diagram of the picture group is determined, the optical flow diagram can be input into the living body detection model trained in the first embodiment, and whether the living body detection object is a living body is determined according to the output data of the living body detection model. Wherein, judging whether the living body detection object is a living body according to the output data of the living body detection model includes: judging whether the living body detection object is a living body according to the confidence level output by the living body detection model (ie, the output data includes the confidence level).

In this embodiment, after the optical flow map is input into the above-mentioned living body detection model, this embodiment further includes:

Let the living body detection model extract the image features of the optical flow map, and calculate the softmax on the image features, so that the living body detection model can determine and output the confidence, that is, input the optical flow map into the above living body detection model, The living body detection model extracts the image features of the optical flow map, calculates softmax on the image features, determines and outputs the confidence level, and the confidence level represents the probability that the living body detection object is a living body.

From the above, it can be determined whether or not the object of living body detection is a living body.

In this embodiment, the picture group contains face pictures at different distances. If the face pictures in the picture group are from a living body, different perspective effects will be formed due to the different distances between the living body and the shooting device. The flow map can accurately represent this perspective effect. The living body detection model in this embodiment can perform live body detection based on the above-mentioned perspective effect. Inputting the optical flow map into the living body detection model can accurately detect whether there is a living body. The living body detection effect is good and the efficiency is high. In this embodiment, the facial image can be aligned in two stages to generate an optical flow map, so that the generated optical flow map (for the above perspective effect) has a better performance and improves the effect and efficiency of living body detection. In this embodiment, the shooting distance of the living body detection object is adapted and prompted to make the living body detection object move, so as to obtain the special optical flow formed by the living body portrait during the movement process, and improve the living body detection effect and efficiency.

The third embodiment of the present specification provides a method for living body detection, which can be regarded as an application of the living body detection model obtained in the first embodiment. The execution body of this embodiment may be a terminal (including but not limited to a mobile phone, a computer, a pad, etc.) or a server, or a corresponding living body detection platform or system or operating system, etc., that is, the execution body may be various, and can be set as required , use or transform the execution body. In addition, there may also be a third-party application program to assist the executive body to execute this embodiment. For example, as shown in FIG. 1 , the living body detection method in this embodiment may be executed by a server, and an application program (corresponding to the server) may be installed on a terminal (held by the user), and the terminal or application program is associated with Data transmission can be performed between servers, data collection, input or output, or page or information display (to the user) can be performed through a terminal or an application program, thereby assisting the server to execute the living body detection method in this embodiment. The following description takes the execution subject as the photographing device as an example.

As shown in FIG. 8 , the live detection method provided in this embodiment includes:

S301: (execution subject) collect a picture group including at least two face pictures at different distances, the face pictures including the face region of the living body detection object;

In this embodiment, a picture group including at least two face pictures at different distances (equivalent to different shooting distances) may be collected. The method described in 2.1 or 2.2 or 2.3 below may be used to obtain a picture group including at least two face pictures with different distances (this embodiment is not limited to the method described in 2.1 or 2.2 or 2.3):

2.1. Video decomposition

In this embodiment, a shooting device (which may include but is not limited to a camera) can shoot a video or picture for a living body detection object (according to the different living body detection results, the living body detection object here may be a living body or a non-living body video or picture, referred to as "" "Non-living", such as remake or generated videos or pictures, the same below). Specifically, the photographing device may start photographing after detecting the living body detection object.

The execution subject may decompose the captured video to obtain several frames of images, and select at least two face pictures at different distances from the several frames of images to form a picture group. Whether the distances of each frame of pictures are different can be determined by using the method provided in the first embodiment, the same below.

2.2. Collect pictures

In this embodiment, the photographing device may collect a face picture for the living body detection object, and each time a face picture for the living body detection object is taken, it releases prompt information, and the prompt information is used to prompt the living body detection object to approach Or stay away from the photographing device, for example, the prompt message may be "please move closer to or stay away from the camera". Wherein, the photographing device may issue the prompt information in the form of text or audio, the same as 1.2.

It can be expected that if the living body detection object is a living body, after receiving (for example, seeing or hearing, the same below) the prompt information, it will approach or move away from the photographing device; if the living body detection object is a non-living body, it should be If the holder of the object is placed in front of the photographing device for photographing, if the holder of the living body detection object receives the prompt information, the owner of the living body detection object will also move the living body detection object closer to or away from the photographing device.

After the photographing device releases the prompt information, it will photograph the next face picture and send it to the execution subject.

The photographing device repeats the above process, that is, "take the previous face picture - release the prompt information - take the next face picture", so that a series of face pictures can be collected, and at least two face pictures with different distances can be selected from them. to form a picture group.

2.3. Receive pictures and judge the distance

In this embodiment, the photographing device can collect a face picture for the living body detection object, and each time a face picture for the living body detection object is taken, the ratio of the region of interest in the face picture to the picture (same as the The first embodiment) determines prompt information, and publishes the prompt information, the prompt information is used to prompt the living body detection object to approach or move away from the photographing device.

Specifically, a threshold can be set. After each face image is collected, the ratio of the region of interest in the face image to the image is determined. If the ratio is greater than (or greater than or equal to) the threshold, it indicates that the living body detection object If the distance is close to the shooting device, the prompt information can be used to prompt the living body detection object to stay away from the shooting device. For example, the prompt information can be “Please stay away from the camera”; if the ratio is less than or equal to (or less than) the threshold, it indicates that the living body detection object If the object is far away from the photographing device, the prompt information may be used to prompt the living body detection object to approach the photographing device, for example, the prompt information may be "please approach the camera".

After the photographing device releases the prompt information, it will photograph the next face picture and send it to the execution subject.

The photographing device repeats the above process, that is, "take the previous face picture - release the prompt information - take the next face picture", so that a series of face pictures can be collected, and at least two face pictures with different distances can be selected from them. to form a picture group.

In particular, in 1.2 or 1.3, after the photographing device issues the prompt information, it can detect whether the living body detection object has moved, and only after the movement has occurred, the next face picture can be taken. If the living body detection object does not move after the prompt information is issued, the prompt information is continuously issued to prevent the non-living body holder from moving the living body detection object and evading the living body detection when the living body detection object is a non-living body.

In 1.2, the prompt information does not limit the living body detection object to be close to or away from the shooting device. In 1.3 (by proportion), the distance between the living body detection object and the shooting device is judged, and then the prompt information limits the living body detection object to approach or stay away from the shooting device, so that the facial image The face area in the image is not too large or too small, making the face image better. In 1.3, if the living body detection object is close to the shooting device, it can be prompted to stay away from the shooting device, otherwise, it can be prompted to approach the shooting device, that is, it can adapt to the distance between the living body detection object and the shooting device, and improve the quality of face pictures.

In 1.2 and 1.3, if the execution subject has received a sufficient number of face images that satisfy different distances, it may instruct the image capture to no longer capture face images.

The face images in the above-mentioned group of pictures should satisfy the (complete) face region containing the living body detection object, even if the living body detection object is non-living body, it can be foreseen that it also contains the face region for use in the living body detection.

In this embodiment, it can be determined whether the distance between any two facial images is the same or different according to the ratio of the region of interest in each facial image to the facial image. Refer to the content in the first embodiment. .

S303: Send the picture group to the living body detection end, so that the living body detection end determines the optical flow graph corresponding to the picture group; and, input the optical flow graph into the living body detection model obtained in the first embodiment , judging whether the living body detection object is a living body according to the output data of the living body detection model.

The living body detection end may be the execution body of the second embodiment, or may be another body. The content here corresponds to the content of the second embodiment, and will not be repeated here.

For the content not described in detail in this embodiment, reference may be made to the first or second embodiment.

In this embodiment, the picture group contains face pictures at different distances. If the face pictures in the picture group are from a living body, different perspective effects will be formed due to the different distances between the living body and the shooting device. The flow map can accurately represent this perspective effect. The living body detection model in this embodiment can perform live body detection based on the above-mentioned perspective effect. Inputting the optical flow map into the living body detection model can accurately detect whether there is a living body. The living body detection effect is good and the efficiency is high. In this embodiment, the facial image can be aligned in two stages to generate an optical flow map, so that the generated optical flow map (for the above perspective effect) has a better performance and improves the effect and efficiency of living body detection. In this embodiment, the shooting distance of the living body detection object is adapted and prompted to make the living body detection object move, so as to obtain the special optical flow formed by the living body portrait during the movement process, and improve the living body detection effect and efficiency.

As shown in FIG. 9 , a fourth embodiment of the present specification provides an apparatus for training a living body detection model, including:

The sample module 401 is used to obtain a sample set for model training, the samples in the sample set include positive samples and negative samples, and each sample includes at least two face pictures with different distances;

The optical flow module 403 is used to determine the optical flow map corresponding to each sample;

The training module 405 is configured to use the optical flow graph corresponding to each sample to train the living body detection model.

Optionally, determining the optical flow map corresponding to each sample includes:

For any sample, align the face pictures included in the sample, and use the aligned face pictures of the sample to determine the optical flow map corresponding to the sample.

Optionally, for any sample, aligning the face pictures included in the sample includes:

For any sample, perform the first-stage alignment of each facial image included in the sample according to the facial key points in each facial image included in the sample;

According to the machine vision feature of each face picture included in the sample, a second-stage alignment is performed on each face picture of the sample after the first alignment.

Optionally, the machine vision features include at least one of SIFT features, HOG features, SURF features, ORB features, LBP features, and HAAR features.

Optionally, determining the optical flow map corresponding to the sample by using the aligned face pictures of the sample includes:

Calculate the optical flow of each face image after the sample is aligned;

An optical flow map corresponding to the sample is generated according to the optical flow of each face image after the sample is aligned.

Optionally, the sample module 401 determines whether the distance between any two facial pictures is the same or different according to the ratio of the region of interest in each facial picture to the facial picture.

Optionally, using the optical flow graph corresponding to each sample to train the living body detection model includes:

A classification model is trained using the optical flow graph corresponding to each sample to obtain a living body detection model.

As shown in FIG. 10 , a fifth embodiment of the present specification provides a living body detection device, and the device is configured with the living body detection model obtained in the first embodiment, including:

A picture module 501, configured to obtain a picture group including at least two face pictures of different distances, the face pictures including the face region of the living body detection object;

An optical flow module 503, configured to determine an optical flow map corresponding to the picture group;

The detection module 505 is configured to input the optical flow map into the living body detection model, and determine whether the living body detection object is a living body according to the output data of the living body detection model.

Optionally, acquiring a picture group including at least two face pictures with different distances includes:

Receive the shooting video for the living body detection object sent by the shooting device, decompose the shooting video to obtain several frames of images, and select at least two face pictures with different distances from the several frame images to form a picture group;

or,

Receive the last face picture sent by the photographing device, and send an instruction to the photographing device, so that the photographing device issues prompt information, the prompt information is used to prompt the living body detection object to approach or move away from the photographing device; receiving the next face picture sent by the photographing device;

Selecting at least two face pictures at different distances from the received face pictures to form a picture group;

or,

Receive the last face picture sent by the shooting device, determine the ratio of the region of interest in the last face picture to the face picture, and send prompt information to the shooting device, so that the shooting device Publishing the prompt information, where the prompt information is used to prompt the living body detection object to approach or move away from the photographing device; receive the next face picture sent by the photographing device;

At least two face pictures at different distances are selected from each of the received face pictures to form a picture group.

Optionally, the picture module 501 determines whether the distance between any two facial pictures is the same or different according to the ratio of the region of interest in each facial picture to the facial picture.

Optionally, determining the optical flow map corresponding to the picture group includes:

Align each of the face pictures, and use the aligned face pictures to determine an optical flow map corresponding to the picture group.

Optionally, aligning each of the face pictures includes:

Perform the first-stage alignment on each of the facial pictures according to the facial key points in each of the facial pictures;

According to the machine vision feature of each of the face pictures, a second-stage alignment is performed on each of the face pictures after the first alignment.

Optionally, the machine vision features include at least one of SIFT features, HOG features, SURF features, ORB features, LBP features, and HAAR features.

Optionally, using the aligned face pictures to determine the optical flow map corresponding to the picture group includes:

calculating the optical flow of each of the face pictures;

An optical flow map corresponding to the picture group is generated according to the optical flow of each of the face pictures.

Optionally, judging whether the living body detection object is a living body according to the output data of the living body detection model includes:

Whether the living body detection object is a living body is determined according to the confidence level output by the living body detection model.

Optionally, the output data includes confidence; after inputting the optical flow graph into the living body detection model, the detection module 505 is further configured to:

The living body detection model is made to extract image features of the optical flow map, and a softmax is calculated on the image features, so that the living body detection model can determine and output the confidence level.

As shown in FIG. 11 , a sixth embodiment of the present specification provides a living body detection device, including:

The picture module 601 is configured to collect a picture group including at least two face pictures at different distances, the face pictures including the face region of the living body detection object;

The sending module 603 is configured to send the picture group to the living body detection end, so that the living body detection end determines the optical flow map corresponding to the picture group; The living body detection model is used to determine whether the living body detection object is a living body according to the output data of the living body detection model.

Optionally, acquiring a picture group including at least two face pictures with different distances includes:

Collecting a shooting video for a living body detection object, decomposing the shooting video to obtain several frames of images, and selecting at least two face pictures with different distances from the several frames of images to form a picture group;

or,

After collecting the last face picture, release prompt information, where the prompt information is used to prompt the living body detection object to approach or stay away from the shooting device;

If it is judged that the living body detection object has moved after collecting the last face picture, collect the next face picture;

Select at least two face pictures at different distances from the collected face pictures to form a picture group;

or,

After collecting the last face picture, determine and publish prompt information according to the ratio of the region of interest in the last face picture to the face picture, and the prompt information is used to prompt the living body detection object to approach or away from said photographing equipment;

If it is judged that the living body detection object has moved after collecting the last face picture, collect the next face picture;

At least two face pictures at different distances are selected from the collected face pictures to form a picture group.

A seventh embodiment of the present specification provides a device for training a living body detection model, including:

at least one processor;

as well as,

a memory communicatively coupled to the at least one processor;

in,

The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the living body detection model training described in the first embodiment method.

An eighth embodiment of the present specification provides a device for training a living body detection model, including:

at least one processor;

as well as,

a memory communicatively coupled to the at least one processor;

in,

The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the living body detection method of the second embodiment.

The ninth embodiment of the present specification provides a living detection model training device, including:

at least one processor;

as well as,

a memory communicatively coupled to the at least one processor;

in,

The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the living body detection method of the third embodiment.

A tenth embodiment of the present specification provides a computer-readable storage medium, where the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, implement the method described in the first embodiment. Liveness detection model training method.

An eleventh embodiment of the present specification provides a computer-readable storage medium, where the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, implement the description in the second embodiment method of live detection.

A twelfth embodiment of the present specification provides a computer-readable storage medium, where the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, implements the third embodiment. method of live detection.

The above embodiments may be used in combination, and modules with the same name may be the same or different between different embodiments.

While the foregoing has described specific embodiments of this specification, other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that in the embodiments and still achieve desirable results. Additionally, the processes depicted in the figures do not necessarily have to follow the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Each embodiment in this specification is described in a progressive manner, and the same and similar parts between the various embodiments may be referred to each other, and each embodiment focuses on the differences from other embodiments. Especially, for the embodiments of the apparatus, equipment, and non-volatile computer-readable storage medium, since they are basically similar to the method embodiments, the description is relatively simple.

The apparatuses, devices, and non-volatile computer-readable storage media provided in the embodiments of this specification correspond to the methods. Therefore, the apparatuses, devices, and non-volatile computer storage media also have beneficial technical effects similar to those of the corresponding methods. The beneficial technical effects of the method have been described in detail above, therefore, the beneficial technical effects of the corresponding apparatus, equipment, and non-volatile computer storage medium will not be repeated here.

In the 1990s, improvements in a technology could be clearly differentiated between improvements in hardware (eg, improvements to circuit structures such as diodes, transistors, switches, etc.) or improvements in software (improvements in method flow). However, with the development of technology, the improvement of many methods and processes today can be regarded as a direct improvement of the hardware circuit structure. Designers almost get the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that the improvement of a method flow cannot be realized by hardware entity modules. For example, a Programmable Logic Device (PLD) (eg, Field Programmable Gate Array (FPGA)) is an integrated circuit whose logic function is determined by user programming of the device. It is programmed by the designer to "integrate" a digital system on a PLD without having to ask the chip manufacturer to design and manufacture a dedicated integrated circuit chip. And, instead of making integrated circuit chips by hand, these days, much of this programming is done using software called a "logic compiler", which is similar to the software compiler used in program development and writing, but before compiling The original code also has to be written in a specific programming language, which is called Hardware Description Language (HDL), and there is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language) , AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (RubyHardware Description Language), etc. The most commonly used are VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. It should also be clear to those skilled in the art that a hardware circuit for implementing the logic method process can be easily obtained by simply programming the method process in the above-mentioned several hardware description languages and programming it into the integrated circuit.

The controller may be implemented in any suitable manner, for example, the controller may take the form of eg a microprocessor or processor and a computer readable medium storing computer readable program code (eg software or firmware) executable by the (micro)processor , logic gates, switches, application specific integrated circuits (ASICs), programmable logic controllers and embedded microcontrollers, examples of controllers include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20 and Silicon Labs C8051F320, the memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that, in addition to implementing the controller in the form of pure computer-readable program code, the controller can be implemented as logic gates, switches, application-specific integrated circuits, programmable logic controllers and embedded devices by logically programming the method steps. The same function can be realized in the form of a microcontroller, etc. Therefore, such a controller can be regarded as a hardware component, and the devices included therein for realizing various functions can also be regarded as a structure within the hardware component. Or even, the means for implementing various functions can be regarded as both a software module implementing a method and a structure within a hardware component.

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

For the convenience of description, when describing the above device, the functions are divided into various units and described respectively. Of course, when implementing this specification, the functions of each unit may be implemented in one or more software and/or hardware.

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

The specification is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the specification. It will be understood that each flow and/or block in the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the 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 flow or flow of a flowchart and/or a block or blocks of a block diagram.

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 The apparatus implements the functions specified in the flow or flow of the flowcharts and/or the block or blocks of the block diagrams.

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 flowcharts and/or the block or blocks of the block diagrams.

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

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

Computer-readable media includes both persistent and non-permanent, removable and non-removable media, and storage of information may be implemented by any method or technology. Information may be computer readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Flash Memory or other memory technology, Compact Disc Read Only Memory (CD-ROM), Digital Versatile Disc (DVD) or other optical storage, Magnetic tape cartridges, tape-based disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, excludes transitory computer-readable media, such as modulated data signals and carrier waves.

It should also be noted that the terms "comprising", "comprising" or any other variation thereof are intended to encompass a non-exclusive inclusion such that a process, method, article or device comprising a series of elements includes not only those elements, but also Other elements not expressly listed, or which are inherent to such a process, method, article of manufacture, or apparatus are also included. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in the process, method, article of manufacture or device that includes the element.

This specification may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The specification can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including storage devices.

Each embodiment in this specification is described in a progressive manner, and the same and similar parts between the various embodiments may be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and for related parts, please refer to the partial descriptions of the method embodiments.

The above descriptions are merely embodiments of the present specification, and are not intended to limit the present application. Various modifications and variations of this application are possible for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this application shall be included within the scope of the claims of this application.

Claims (27)

1. A method of living body detection model training, comprising:

obtaining a sample set for model training, wherein samples in the sample set comprise positive samples and negative samples, and each sample comprises at least two face pictures with different distances;

determining an optical flow graph corresponding to each sample;

and training a living body detection model by using the optical flow graph corresponding to each sample.

2. The method of claim 1, determining the optical flow graph corresponding to each sample comprises:

and for any sample, aligning each face picture contained in the sample, and determining an optical flow graph corresponding to the sample by using each face picture after the sample is aligned.

3. The method of claim 2, wherein for any sample, aligning the face pictures included in the sample comprises:

carrying out first-stage alignment on each face picture contained in any sample according to the face key point in each face picture contained in the sample;

and according to the machine vision characteristics of each face picture contained in the sample, carrying out second-stage alignment on each face picture of the sample after the first alignment.

4. The method of claim 3, the machine vision features comprising at least one of SIFT features, HOG features, SURF features, ORB features, LBP features, HAAR features.

5. The method of claim 2, wherein determining the optical flow graph corresponding to the sample using the sample-aligned face pictures comprises:

calculating the optical flow of each face picture after the sample is aligned;

and generating an optical flow graph corresponding to the sample according to the optical flow of each face picture after the sample is aligned.

6. The method of claim 1, wherein the distance between any two face pictures is determined according to the proportion of the interest region in each face picture to the face picture.

7. The method of claim 1, wherein training a liveness detection model using the optical flow maps for each sample comprises:

and training a classification model by using the optical flow graph corresponding to each sample to obtain a living body detection model.

8. A method of in vivo detection comprising:

acquiring a picture group comprising at least two face pictures with different distances, wherein the face pictures comprise face areas of living body detection objects;

determining an optical flow graph corresponding to the picture group;

inputting the optical flow diagram into the living body detection model obtained in any one of claims 1 to 7, and determining whether or not the living body detection object is a living body from output data of the living body detection model.

9. The method of claim 8, wherein obtaining a group of pictures containing at least two different distance facial pictures comprises:

receiving a shooting video aiming at a living body detection object and sent by shooting equipment, decomposing the shooting video to obtain a plurality of frame images, and selecting at least two face pictures with different distances from the plurality of frame images to form a picture group;

or,

receiving a last face picture sent by a shooting device, and sending an instruction to the shooting device so as to enable the shooting device to issue prompt information, wherein the prompt information is used for prompting that the living body detection object is close to or far away from the shooting device; receiving the next face picture sent by the shooting equipment;

selecting at least two face pictures with different distances from each received face picture to form a picture group;

or,

receiving a previous face picture sent by a shooting device, determining according to the proportion of an interest region in the previous face picture and the previous face picture, and sending prompt information to the shooting device so that the shooting device can issue the prompt information, wherein the prompt information is used for prompting that the living body detection object is close to or far away from the shooting device; receiving the next face picture sent by the shooting equipment;

at least two different distance face pictures are selected from each received face picture to form a picture group.

10. The method of claim 8, wherein the distance between any two face pictures is determined according to the proportion of the interest region in each face picture to the face picture.

11. The method of claim 8, determining an optical flow map corresponding to the group of pictures comprises:

aligning each face picture, and determining an optical flow graph corresponding to the picture group by using each aligned face picture.

12. The method of claim 11, aligning each of the facial pictures comprises:

performing first-stage alignment on each face picture according to the face key points in each face picture;

and performing second-stage alignment on each face picture subjected to the first alignment according to the machine vision characteristics of each face picture.

13. The method of claim 12, the machine vision features comprising at least one of SIFT features, HOG features, SURF features, ORB features, LBP features, HAAR features.

14. The method of claim 11, wherein determining an optical flow map corresponding to the group of pictures using the aligned each of the facial pictures comprises:

calculating optical flow of each face picture;

and generating an optical flow graph corresponding to the picture group according to the optical flow of each face picture.

15. The method of claim 8, determining whether the living body detection object is a living body according to the output data of the living body detection model comprises:

and judging whether the living body detection object is a living body according to the confidence coefficient output by the living body detection model.

16. The method of claim 8, the output data comprising a confidence level; after inputting the optical flow diagram into the in-vivo detection model obtained according to any one of claims 1 to 7, the method further comprises:

enabling the living body detection model to extract image features of the light flow graph, and calculating softmax for the image features so as to enable the living body detection model to determine and output the confidence coefficient.

17. A living body detection method is applied to shooting equipment and comprises the following steps:

acquiring a picture group comprising at least two face pictures with different distances, wherein the face pictures comprise face regions of living body detection objects;

sending the picture group to a living body detection end so that the living body detection end determines an optical flow graph corresponding to the picture group; and inputting the optical flow diagram into the living body detection model obtained in any one of claims 1 to 7, and determining whether or not the living body detection object is a living body based on output data of the living body detection model.

18. The method of claim 17, wherein obtaining a group of pictures containing at least two different distance facial pictures comprises:

acquiring a shooting video aiming at a living body detection object, decomposing the shooting video to acquire a plurality of frame images, and selecting at least two face pictures with different distances from the plurality of frame images to form a picture group;

or,

after the last face picture is collected, issuing prompt information, wherein the prompt information is used for prompting that the living body detection object approaches to or leaves away from the shooting equipment;

if the living body detection object is judged to move after the previous face picture is collected, collecting the next face picture;

selecting at least two face pictures with different distances from each collected face picture to form a picture group;

or,

after the last face picture is collected, determining and issuing prompt information according to the proportion of the interest area in the last face picture and the face picture, wherein the prompt information is used for prompting the living body detection object to approach or leave the shooting equipment;

if the living body detection object is judged to move after the previous face picture is collected, collecting the next face picture;

at least two face pictures with different distances are selected from each acquired face picture to form a picture group.

19. A living body detection model training apparatus, comprising:

the system comprises a sample module, a face recognition module and a face recognition module, wherein the sample module is used for obtaining a sample set used for model training, samples in the sample set comprise positive samples and negative samples, and each sample comprises at least two face pictures with different distances;

the optical flow module is used for determining an optical flow graph corresponding to each sample;

and the training module is used for training a living body detection model by using the optical flow graph corresponding to each sample.

20. A living body test apparatus provided with the living body test model obtained in any one of claims 1 to 7, comprising:

the image module is used for acquiring an image group comprising at least two face images with different distances, wherein the face images comprise face areas of living body detection objects;

the optical flow module is used for determining an optical flow graph corresponding to the picture group;

and the detection module is used for inputting the optical flow diagram into the living body detection model and judging whether the living body detection object is a living body according to output data of the living body detection model.

21. A living body detection apparatus comprising:

the image module is used for acquiring an image group comprising at least two face images with different distances, wherein the face images comprise face areas of living body detection objects;

the sending module is used for sending the picture group to a living body detection end so that the living body detection end determines an optical flow graph corresponding to the picture group; and inputting the optical flow diagram into the living body detection model obtained in any one of claims 1 to 7, and determining whether or not the living body detection object is a living body based on output data of the living body detection model.

22. A living body detection model training apparatus, comprising:

at least one processor;

and the number of the first and second groups,

a memory communicatively coupled to the at least one processor;

wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the liveness detection model training method of any one of claims 1 to 7.

23. A living body examination apparatus comprising:

at least one processor;

and the number of the first and second groups,

a memory communicatively coupled to the at least one processor;

wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the liveness detection method of any one of claims 8 to 16.

24. A living body examination apparatus comprising:

at least one processor;

and the number of the first and second groups,

a memory communicatively coupled to the at least one processor;

wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the liveness detection method of claim 17 or 18.

25. A computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the in-vivo detection model training method of any one of claims 1 to 7.

26. A computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the liveness detection method of any one of claims 8 to 16.

27. A computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the liveness detection method of claim 17 or 18.

Images