WO2019071664A1

WO2019071664A1 - Human face recognition method and apparatus combined with depth information, and storage medium

Info

Publication number: WO2019071664A1
Application number: PCT/CN2017/108804
Authority: WO
Inventors: 戴磊
Original assignee: 平安科技（深圳）有限公司
Priority date: 2017-10-09
Filing date: 2017-10-31
Publication date: 2019-04-18
Also published as: CN107844744A

Abstract

Disclosed is a human face recognition method combined with depth information. The method comprises: establishing a human face image sample library, wherein a human face ID, a human face color image and a human face depth image match each other in same; constructing and training a human face classification and recognition model to obtain a feature vector of a human face sample image; acquiring a human face image of a target to be recognized, the human face image comprising a human face color image of the target to be recognized and a human face depth image of the target to be recognized; inputting the human face image of the target to be recognized into the human face classification and recognition model, and extracting a feature vector of the human face image of the target to be recognized; and determining, according to the feature vector, a human face ID corresponding to the human face image of the target to be recognized. In combination with depth information about a human face, by means of calculating the vector distance between feature vectors of a human face color image and depth image of a target to be recognized and a feature vector of a human face sample image, the present application realizes precise recognition of the human face. Further provided are an electronic apparatus and a computer-readable storage medium.

Description

Face recognition method, device and storage medium combined with depth information

This patent application is based on the Chinese Patent Application No. 201710930240.5, entitled "Face Recognition Method, Apparatus and Storage Medium Combining Depth Information", which is filed on October 9, 2017, and claims priority.

Technical field

The present application relates to the field of face recognition technologies, and in particular, to a face recognition method, apparatus, and storage medium that combine depth information.

Background technique

Face recognition is a non-contact biometric recognition technology based on human facial feature information for identity authentication. With the rapid development of computer and network technologies, face recognition technology has been widely used in many important industries and fields such as smart access control, public safety, entertainment, and military.

Currently, face recognition is commonly used in 2D image-based Convolutional Neural Networks (CNN) for regression. Since there is no depth information in 2D images, the existing face recognition system can distinguish similar faces. It is generally inadequate and has poor adaptability to texture changes on the same face, such as face painting, tattoos, or changes in lighting.

Summary of the invention

The present invention provides a face recognition method, device and storage medium combining depth information, and the main purpose thereof is to realize accurate recognition of a human face by combining a face color image and a depth image.

To achieve the above objective, the present application provides a face recognition method, including:

The sample library establishing step: acquiring 2N face images, including N face color images and corresponding N face depth images, preprocessing the face images, and using the preprocessed face images as face samples An image, a face image sample library in which a face ID, a face color image, and a face depth image are matched with each other, wherein N is an integer greater than 2, and the face depth image includes each point of the person's face and the image acquisition unit Distance information;

a training step of the model: training the convolutional neural network with the matching face sample images to obtain a feature vector of the face classification recognition model and the face sample image;

Target face image acquisition step: acquiring a target face image to be recognized, including a face color image and a corresponding face depth image;

a target recognition step: inputting the target face image to be recognized into the face classification recognition model, extracting a feature vector of the target face image to be recognized, and searching and searching the face image sample library according to the feature vector A face sample image to which the target face image to be recognized matches, and a face ID of the target face image to be recognized is determined according to the face sample image.

Preferably, the preprocessing in the sample library establishing step comprises performing a process of removing image noise and correcting a face pose on the face image, and labeling the face image with a face ID.

Preferably, in the target recognition step, according to the feature vector of the target face image to be identified Finding a face sample image matching the target face image to be recognized in the face image sample library includes:

Calculating a vector distance between a feature vector of the target face image to be identified and a feature vector of the face sample image;

A face sample image whose obtained vector distance is the smallest or smaller than the threshold is taken as a face sample image that matches the target face image to be recognized.

Optionally, the vector distance is a cosine distance or an Euclidean distance.

Preferably, the 2N face images in the sample library establishing step are obtained by the following method:

The first photographing step: using a camera to capture a color image of a person appearing in a shooting area within a preset time range and a corresponding depth image;

The first face detecting step: extracting a face image from the color image and the depth image by using a face detection algorithm to obtain N face color images and corresponding N face depth images.

The target face image is obtained by the following method:

a second shooting step: capturing, by the camera, a color image and a depth image of the object to be recognized appearing in the current shooting area;

The second face detecting step is: extracting a face image from the color image and the depth image of the object to be recognized by using a face detection algorithm to obtain a face color image and a face depth image of the object to be recognized.

Optionally, the face recognition detection algorithm is one or more of a geometric feature based algorithm, a local feature analysis algorithm, a feature face algorithm, an elasticity model based algorithm, and a neural network algorithm.

In addition, in order to achieve the above object, the present application further provides an electronic device including an image acquisition unit, a memory, and a processor, the image acquisition unit including a camera having a depth imaging function including face recognition in the memory a program that implements the following steps when executed by the processor:

Preferably, the searching for the face sample image matching the to-be-identified target face image in the face image sample library according to the feature vector of the target face image to be recognized in the target recognition step includes:

Optionally, the vector distance is a cosine distance or an Euclidean distance.

The target face image is obtained by the following method:

In addition, in order to achieve the above object, the present application further provides a computer readable storage medium including a face image sample library, a face classification recognition model, and a face recognition program, the face recognition program Any of the steps of the face recognition method of combining depth information as described above when executed by the processor.

The face recognition method, the electronic device and the computer readable storage medium combined with the depth information proposed by the present application, by constructing and training the face classification recognition model by inputting the face color image and the face depth image of the target to be recognized in real time Extracting a feature vector of the target face image to be recognized, calculating a vector distance of the feature vector of the target face image to be recognized and the feature vector of the face sample image, and searching and searching in the face image sample library according to the vector distance The face sample image matched by the target face image to be identified. Since the face classification recognition model is trained and the feature vector of the extracted face image is applied with the face depth image, and the face depth image includes the distance information of each point of the person's face and the image acquisition unit, the present application can achieve more on the face. Accurate identification, especially when the planar features of the face are highly similar, while stereo features such as nose height, orbital depth, and tibia height are different In this case, the face recognition accuracy can be significantly improved.

DRAWINGS

1 is a schematic diagram of an operating environment of a preferred embodiment of an electronic device of the present application;

2 is a functional block diagram of the face recognition program of FIG. 1;

3 is a flowchart of a face recognition method combining depth information according to the present application;

4 is a flow chart of a preferred embodiment of a face recognition method incorporating depth information in the present application.

The implementation, functional features and advantages of the present application will be further described with reference to the accompanying drawings.

Detailed ways

It is understood that the specific embodiments described herein are merely illustrative of the application and are not intended to be limiting.

1 is a schematic diagram of an operating environment of a preferred embodiment of an electronic device of the present application.

The electronic device 1 may be a terminal device having an imaging and computing function, such as a server, a smart phone, a tablet computer, a portable computer, or a desktop computer.

Referring to FIG. 1, the electronic device 1 includes an image acquisition unit 11, a memory 12, a processor 13, a network interface 14, and a communication bus 15. The image obtaining unit 11 can be installed in a specific place, such as an office place and a monitoring area, and real-time images are taken in real time for the target entering the specific place, and the captured real-time image is transmitted to the processor 13 through the network. Network interface 14 may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface). Communication bus 15 is used to implement connection communication between these components.

Memory 12 includes at least one type of readable storage medium. The at least one type of readable storage medium may be a non-volatile storage medium such as a flash memory, a hard disk, a multimedia card, a card type memory, or the like. In some embodiments, the readable storage medium may be an internal storage unit of the electronic device 1, such as a hard disk of the electronic device 1. In other embodiments, the readable storage medium may also be an external memory 11 of the electronic device 1, such as a plug-in hard disk equipped on the electronic device 1, a smart memory card (SMC). , Secure Digital (SD) card, Flash Card, etc.

In this embodiment, the readable storage medium of the memory 12 is generally used to store the face recognition program 10 installed on the electronic device 1, the face image sample library, and the face classification recognition model constructed and trained. . The memory 12 can also be used to temporarily store data that has been output or is about to be output.

The processor 13, in some embodiments, may be a Central Processing Unit (CPU), microprocessor or other data processing chip for running program code or processing data stored in the memory 12, such as executing a human face. Identify the program 10 and so on.

Figure 1 shows only the electronic device 1 with the components 11-15 and the face recognition program 10, but it should be understood that not all of the illustrated components are required to be implemented, and alternative implementations may be more or less s component.

Optionally, the electronic device 1 may further include a user interface, and the user interface may include an input unit such as a keyboard, a voice input device such as a microphone, a device with a voice recognition function, a voice output device such as an audio, a headphone, and the like. Optionally, the user interface may also include a standard wired interface and a wireless interface.

The electronic device 1 may also include a display, which may also be referred to as a display screen or display unit as appropriate. In some embodiments, it may be an LED display, a liquid crystal display, a touch liquid crystal display, and an Organic Light-Emitting Diode (OLED) display. The display is used to display information processed in the electronic device 1 and a user interface for displaying visualizations.

The electronic device 1 also includes a touch sensor. The area provided by the touch sensor for the user to perform a touch operation is referred to as a touch area. Further, the touch sensor described herein may be a resistive touch sensor, a capacitive touch sensor, or the like. Moreover, the touch sensor includes not only a contact type touch sensor but also a proximity type touch sensor or the like. Furthermore, the touch sensor may be a single sensor or a plurality of sensors arranged, for example, in an array.

In addition, the area of the display of the electronic device 1 may be the same as or different from the area of the touch sensor. Optionally, a display is stacked with the touch sensor to form a touch display. The device detects a user-triggered touch operation based on a touch screen display.

The electronic device 1 may further include a radio frequency (RF) circuit, a sensor, an audio circuit, and the like, and details are not described herein.

In the operating environment diagram of the preferred embodiment of the electronic device 1 shown in FIG. 1, the memory 12 as a readable storage medium may include an operating system, a face recognition program 10, a face image sample library, and build and train. Good face classification recognition model. When the processor 13 executes the face recognition program 10 stored in the memory 12, the following steps are implemented:

The sample library establishing step: acquiring 2N face images, including N face color images and corresponding N face depth images, preprocessing the face images, and using the preprocessed face images as face samples An image, a face image sample library in which a face ID, a face color image, and a face depth image are matched with each other, wherein N is an integer greater than 2, and the face depth image includes each point of the person's face and the image acquisition unit 11 Distance information;

a target recognition step: inputting the target face image to be recognized into the face classification recognition model, extracting a feature vector of the target face image to be recognized, and searching and searching the face image sample library according to the feature vector A face sample image to which the target face image to be recognized matches, and a face ID of the target face image to be recognized is determined according to the face sample image. For the specific principle, please refer to the following FIG. 2 for the function block diagram of the face recognition program 10 and FIG. 3 for the face recognition method for combining the depth information. Introduction to the flow chart.

Referring to FIG. 2, it is a functional block diagram of the face recognition program 10 of FIG. In the present embodiment, the face recognition program 10 is divided into a plurality of modules, which are stored in the memory 12 and executed by the processor 13 to complete the present application. A module as referred to in this application refers to a series of computer program instructions that are capable of performing a particular function.

The face recognition program 10 can be divided into: an acquisition module 110, a learning module 120, an extraction module 130, and an identification module 140.

The obtaining module 110 is configured to acquire a color image and a depth image of the person captured by the image acquiring unit 11, extract a face image therefrom and perform preprocessing on the extracted face image by using a face detection algorithm. For example, as shown in FIG. 3, the face image sample library may be established by the acquisition module 110 to obtain a target face color image and a depth image to be recognized.

The image acquisition unit 11 may be a camera integrated with a common camera function and a depth camera function, or may include two cameras: an ordinary camera that captures a color image, and a depth camera that captures a depth image (eg, Time of Flight). Camera). The color image can be described by a variety of color spaces, including RGB, HSV, HIS, CMY, and the like. Each pixel value in the depth image represents a distance between a certain point in the scene and the camera, and in the present embodiment, represents a distance between a certain point of the face and the image acquisition unit 11. When the image acquisition unit 11 captures a color image and a depth image, the shooting angle and the frequency are set to be the same. For example, the image acquisition unit 11 is adjusted to the direction facing the entrance of the monitoring area (photographing the front image of the person), and setting a color image and a depth image at the same time every fixed time (for example, 2 seconds), and color images taken at the same time and The depth image can establish a one-to-one correspondence by labeling the shooting time. The pre-processing includes removing image noise, correcting a face pose, and animating the face ID for each face color image and depth image. The image noise may be caused by various reasons. For example, due to some deficiencies that may exist in the device itself, some points on the face of the person may not be acquired by the image acquisition unit 11 and the depth portion may be replaced by a value of 0. Output, these zero values can be considered as noise. In this embodiment, the bilinear interpolation algorithm may be used to fill the portion of the pixels that cannot be acquired by the image acquisition unit 11 and then the Gaussian filter is used to remove the noise interference.

The learning module 120 is configured to train the convolutional neural network with a face image sample library in which the face ID, the face color image, and the face depth image match each other, to obtain a face classification recognition model and the face sample image. Feature vector.

In the process of training the convolutional neural network with the face image sample library, each set of matching face color images and depth images are input into a convolutional neural network, and the set of face color images and depth images are extracted. The face feature vector of the corresponding character, the more samples used in the training model and the larger the sample difference, the more accurate the obtained face classification model. As shown in FIG. 3, the feature vectors extracted by the convolutional neural network for each group of face color images and depth images are T1, T2, T3, ..., Tn.

The extraction module 130 is configured to input the target face color image and the depth image to be input into the face classification recognition model, and extract the target face image to be recognized by using the face classification recognition model A feature vector of the image, such as a feature vector T.

The identification module 140 is configured to search, in the face image sample library, the face sample image that matches the target face image to be recognized according to the feature vector, and determine the target face to be recognized according to the face sample image. The face ID of the image.

In this embodiment, the identification module 140 calculates a vector distance between the feature vector of the target face sample image to be recognized and the feature vector of the face sample image, and uses the face sample image with the obtained vector distance minimum or less than the threshold as the context. A face sample image in which the recognition target face image matches is described, and a face ID of the target face image to be recognized is determined according to the face sample image.

For example, as shown in FIG. 3, the identification module 140 calculates the distances D1, D2, D3, ... of the feature vector T of the target face image to be recognized and the feature vectors T1, T2, T3, ..., Tn of the face sample image. Dn, the minimum value is selected from the distances D1, D2, D3, ..., Dn, or all the distance values smaller than the preset threshold are selected, and the face sample image corresponding to the selected distance value is the same to be identified. A face sample image in which the target face image matches, and the face ID of the face sample image is the face ID of the target face image to be recognized. The vector distance may be a cosine distance or an Euclidean distance.

Referring to FIG. 4, it is a flowchart of a preferred embodiment of a face recognition method incorporating depth information according to the present application. The following steps of implementing the face recognition method when the processor 13 of the electronic device 1 executes the face recognition program 10 stored in the memory 12:

Step S10: The acquisition module 110 establishes a face image sample library in which the face ID, the face color image, and the face depth image match each other. The obtaining module 110 acquires a color image and a depth image of a person appearing in a shooting area within a preset time range, and extracts a face image from the face detection algorithm to obtain a face color image and a corresponding face depth image, and the face image is obtained The pre-processing is performed, and the pre-processed face image is used as a face sample image, and a face image sample library in which the face ID, the face color image, and the face depth image match each other is established. The pre-processing includes performing a process of removing image noise and correcting a face pose on the 2N face images, and labeling the face image with the face ID.

In step S20, the face classification recognition model is constructed and trained by the learning module 120 to obtain a feature vector of the face sample image. The learning module 120 trains the convolutional neural network with the face sample images that match each other to obtain a feature vector of the face classification recognition model and the face sample image. When training the face classification recognition model, the more the number of face sample images used and the greater the difference, the more accurate the obtained face classification model.

In step S30, the image of the target face to be identified is acquired by the acquiring module 110, including the target face color image and the face depth image to be identified. The acquiring module 110 acquires a color image and a depth image of the object to be recognized that appear in the current shooting area, and extracts a face image from the color image and the depth image of the object to be identified by using a face detection algorithm to obtain a face color of the target to be identified. Image and face depth image. In this embodiment, the face recognition detection algorithm is one or more of a geometric feature based algorithm, a local feature analysis algorithm, a feature face algorithm, an elasticity model based algorithm, and a neural network algorithm.

In step S40, the target face image to be recognized is input into the face classification recognition model, and the extraction module 130 extracts a feature vector of the target face image to be recognized.

In step S50, the identification module 140 searches the face image sample library for the face sample image matching the target face image to be recognized according to the feature vector of the target face image to be recognized, and determines the face sample image according to the face sample image. The face ID of the target face image to be identified. The recognition module 140 calculates a vector sample distance between the feature vector of the target face image to be recognized and the feature vector of the face sample image, and takes the face sample image with the obtained vector distance minimum or smaller than the threshold as A face sample image in which the target face image matches is identified, and the face ID of the face sample image is the face ID of the target face image to be recognized. The vector distance may be a cosine distance or an Euclidean distance.

The face recognition method proposed in this embodiment combines the face plane information and the depth information, and uses the face classification recognition model to extract the feature color of the face color image and the face depth image of the object to be recognized, according to the feature vector. The face image sample library searches for a face sample image that matches the target face image to be recognized. Since the face classification recognition model is trained, the feature vector of the extracted face image is applied with the face depth image, and the face depth image includes the distance information of each point of the person face and the image acquisition unit 11, the present application can implement the face More accurate recognition, especially when the planar features of the human face are highly similar, and the stereoscopic features, such as the height of the bridge of the nose, the depth of the eye socket, and the height of the tibia, can significantly improve the accuracy of face recognition.

In addition, the embodiment of the present application further provides a computer readable storage medium, which may be a hard disk, a multimedia card, an SD card, a flash memory card, an SMC, a read only memory (ROM), and an erasable programmable Any combination or combination of any one or more of read only memory (EPROM), portable compact disk read only memory (CD-ROM), USB memory, and the like. The computer readable storage medium includes a face image sample library, a constructed and trained face classification recognition model, a face recognition program 10, and the like, and the face recognition program 10 is implemented by the processor 13 as follows operating:

a target face image obtaining step: acquiring a target face image to be recognized, including a face color image of the target to be identified and a corresponding face depth image;

The specific implementation of the computer-readable storage medium of the present application is substantially the same as the above-described embodiment of the face recognition method and the electronic device 1 in combination with the depth information, and details are not described herein again.

It is to be understood that the term "comprises", "comprising", or any other variants thereof, is intended to encompass a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a series of elements includes those elements. It also includes other elements not explicitly listed, or elements that are inherent to such a process, device, item, or method. An element that is defined by the phrase "comprising a ..." does not exclude the presence of additional equivalent elements in the process, the device, the item, or the method that comprises the element.

Through the description of the above embodiments, those skilled in the art can clearly understand that the foregoing embodiment method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be through hardware, but in many cases, the former is better. Implementation. Based on such understanding, portions of the technical solution of the present application that contribute substantially or to the prior art may be embodied in the form of a software product stored in a storage medium as described above, including a number of instructions. To enable a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the methods described in the various embodiments of the present application.

The above is only a preferred embodiment of the present application, and is not intended to limit the scope of the patent application, and the equivalent structure or equivalent process transformations made by the specification and the drawings of the present application, or directly or indirectly applied to other related technical fields. The same is included in the scope of patent protection of this application.

Claims

A face recognition method combining depth information, the method comprising:

The sample library establishing step: acquiring 2N face images, including N face color images and corresponding N face depth images, preprocessing the face images, and using the preprocessed face images as face samples An image, a face image sample library in which a face ID, a face color image, and a face depth image are matched with each other, wherein N is an integer greater than 2;

a training step of the model: training the convolutional neural network with the matching face sample images to obtain a feature vector of the face classification recognition model and the face sample image;

Target face image acquisition step: acquiring a target face image to be recognized, including a face color image and a corresponding face depth image;

a target recognition step: inputting the target face image to be recognized into the face classification recognition model, extracting a feature vector of the target face image to be recognized, and searching and searching the face image sample library according to the feature vector A face sample image to which the target face image to be recognized matches, and a face ID of the target face image to be recognized is determined according to the face sample image.
The face recognition method according to claim 1, wherein the pre-processing comprises performing a process of removing image noise and correcting a face pose on the face image, and labeling the face image with a face ID .
The face recognition method according to claim 1, wherein the 2N face images in the sample library establishing step are obtained by the following method:

The first photographing step: using a camera to capture a color image of a person appearing in a shooting area within a preset time range and a corresponding depth image;

The first face detecting step: extracting a face image from the color image and the depth image by using a face detection algorithm to obtain N face color images and corresponding N face depth images.
The face recognition method according to claim 1, wherein the target face image acquisition step comprises:

a second shooting step: capturing, by the camera, a color image and a depth image of the object to be recognized appearing in the current shooting area;

The second face detecting step is: extracting a face image from the color image and the depth image of the object to be recognized by using a face detection algorithm to obtain a face color image and a face depth image of the object to be recognized.
The face recognition method according to claim 3 or 4, wherein the face recognition detection algorithm is a geometric feature based algorithm, a local feature analysis algorithm, a feature face algorithm, an elastic model based algorithm, a neural network algorithm. One or several of them.
The face recognition method according to claim 1, wherein the target recognition step searches for the target person to be identified in the face image sample database according to the feature vector of the target face image to be recognized. The face sample image matching the face image includes:

Calculating a vector distance between a feature vector of the target face image to be identified and a feature vector of the face sample image;

A face sample image whose obtained vector distance is the smallest or smaller than the threshold is taken as a face sample image that matches the target face image to be recognized.
The face recognition method according to claim 6, wherein the vector distance is a cosine distance or an Euclidean distance.
An electronic device comprising an image acquisition unit, a memory and a processor, wherein the image acquisition unit comprises a camera having a depth camera function, wherein the memory includes a face recognition program, the face recognition program being The processor implements the following steps when it executes:

The sample library establishing step: acquiring 2N face images, including N face color images and corresponding N face depth images, preprocessing the face images, and using the preprocessed face images as face samples An image, a face image sample library in which a face ID, a face color image, and a face depth image are matched with each other, wherein N is an integer greater than 2;

a training step of the model: training the convolutional neural network with the matching face sample images to obtain a feature vector of the face classification recognition model and the face sample image;

Target face image acquisition step: acquiring a target face image to be recognized, including a face color image and a corresponding face depth image;

a target recognition step: inputting the target face image to be recognized into the face classification recognition model, extracting a feature vector of the target face image to be recognized, and searching and searching the face image sample library according to the feature vector A face sample image to which the target face image to be recognized matches, and a face ID of the target face image to be recognized is determined according to the face sample image.
The electronic device according to claim 8, wherein the pre-processing comprises performing a process of removing image noise and correcting a face pose on the face image, and labeling the face image with a face ID.
The electronic device according to claim 8, wherein the 2N face images in the sample library establishing step are obtained by the following method:

The first photographing step: using a camera to capture a color image of a person appearing in a shooting area within a preset time range and a corresponding depth image;

The first face detecting step: extracting a face image from the color image and the depth image by using a face detection algorithm to obtain N face color images and corresponding N face depth images.
The electronic device according to claim 8, wherein said target face image The acquisition steps include:

a second shooting step: capturing, by the camera, a color image and a depth image of the object to be recognized appearing in the current shooting area;

The second face detecting step is: extracting a face image from the color image and the depth image of the object to be recognized by using a face detection algorithm to obtain a face color image and a face depth image of the object to be recognized.
The electronic device according to claim 10 or 11, wherein the face recognition detection algorithm is based on a geometric feature based algorithm, a local feature analysis algorithm, a feature face algorithm, an elastic model based algorithm, and a neural network algorithm One or several.
The electronic device according to claim 8, wherein in the target recognition step, the face image of the object to be recognized is searched for in the face image sample library according to the feature vector of the target face image to be recognized. The matching face sample images include:

Calculating a vector distance between a feature vector of the target face image to be identified and a feature vector of the face sample image;

A face sample image whose obtained vector distance is the smallest or smaller than the threshold is taken as a face sample image that matches the target face image to be recognized.
The electronic device according to claim 13, wherein the vector distance is a cosine distance or an Euclidean distance.
A computer readable storage medium, comprising: a face recognition program, a face image sample library, and a face classification recognition model, wherein when the face recognition program is executed by a processor, Implement the following steps:

The sample library establishing step: acquiring 2N face images, including N face color images and corresponding N face depth images, preprocessing the face images, and using the preprocessed face images as face samples An image, a face image sample library in which a face ID, a face color image, and a face depth image are matched with each other, wherein N is an integer greater than 2;

a training step of the model: training the convolutional neural network with the matching face sample images to obtain a feature vector of the face classification recognition model and the face sample image;

Target face image acquisition step: acquiring a target face image to be recognized, including a face color image and a corresponding face depth image;

a target recognition step: inputting the target face image to be recognized into the face classification recognition model, extracting a feature vector of the target face image to be recognized, and searching and searching the face image sample library according to the feature vector A face sample image to which the target face image to be recognized matches, and a face ID of the target face image to be recognized is determined according to the face sample image.
The computer readable storage medium of claim 1 wherein said pre- The processing includes performing a process of removing image noise and correcting a face pose on the face image, and labeling the face image with a face ID.
The computer readable storage medium according to claim 1, wherein the 2N face images in the sample library establishing step are obtained by the following method:

The first photographing step: using a camera to capture a color image of a person appearing in a shooting area within a preset time range and a corresponding depth image;

The first face detecting step: extracting a face image from the color image and the depth image by using a face detection algorithm to obtain N face color images and corresponding N face depth images.
The computer readable storage medium according to claim 1, wherein the target face image acquisition step comprises:

a second shooting step: capturing, by the camera, a color image and a depth image of the object to be recognized appearing in the current shooting area;

The second face detecting step is: extracting a face image from the color image and the depth image of the object to be recognized by using a face detection algorithm to obtain a face color image and a face depth image of the object to be recognized.
The computer readable storage medium according to claim 1, wherein the face recognition detection algorithm is a geometric feature based algorithm, a local feature analysis algorithm, a feature face algorithm, an elastic model based algorithm, and a neural network algorithm. One or several.
The computer readable storage medium according to claim 1, wherein the target recognition step searches for the target to be recognized in the face image sample library according to the feature vector of the target face image to be recognized. The face sample image matching the face image includes:

Calculating a vector distance between a feature vector of the target face image to be identified and a feature vector of the face sample image;

A face sample image whose obtained vector distance is the smallest or smaller than the threshold is taken as a face sample image that matches the target face image to be recognized.