WO2019029261A1

WO2019029261A1 - Micro-expression recognition method, device and storage medium

Info

Publication number: WO2019029261A1
Application number: PCT/CN2018/090990
Authority: WO
Inventors: 袁晖
Original assignee: 深圳市科迈爱康科技有限公司
Priority date: 2017-08-07
Filing date: 2018-06-13
Publication date: 2019-02-14
Also published as: CN107480622A

Abstract

Disclosed in the present application of the present invention are a micro-expression recognition method, a device and a storage medium. Said method comprises: performing image recognition on a video to be recognized, obtaining a face part in the video to be recognized, and dividing the face part according to pre-set regions; extracting, from the video to be recognized, expression feature information in each pre-set region; comparing the expression feature information with a pre-set micro-expression model, and determining, according to the comparison result, a micro-expression in the video to be recognized.

Description

Micro expression recognition method, device and storage medium

Technical field

The present application relates to the field of communications technologies, and in particular, to a micro-expression recognition method, apparatus, and storage medium.

Background technique

By making some expressions, people express their inner feelings to others. In these different expressions, or in an expression, the face will “leak” other information through the micro-expression. Micro-expressions generally only last for 1/25~1/5 seconds. Although a subconscious micro-expression may last only for a moment, it is easy to expose people's true emotions. Therefore, micro-expression recognition has an extraordinary effect on analyzing people's true mental state. With the rapid development of computer vision, pattern recognition and other disciplines, the automatic recognition technology of micro-expressions is quite mature. The research on automatic recognition of micro-expressions has been greatly developed in recent years, and several standard micros have been established at home and abroad. Expression library.

However, the micro-expression library used in the current micro-expression recognition method is established under the unnatural conditions such as expression suppression, and is quite different from the actual life scene of the people, and cannot accurately reflect the true state of the micro-expression. Therefore, a micro-expression library established by capturing people's micro-expressions in real life state is needed, and an identification method capable of better reflecting the true state of the micro-expressions is determined by the micro-expression library.

Summary of the invention

The main purpose of the present application is to provide a micro-expression recognition method, device and storage medium, which aim to solve the technical problem that the actual situation of the micro-expression cannot be better reflected in the prior art.

To achieve the above objective, the present application provides a micro-expression recognition method, the method comprising the following steps:

Performing image recognition on the recognized video, obtaining a face part in the to-be-identified video, and dividing the face part according to the preset area;

Extracting expression feature information of each preset area from the to-be-identified video;

Comparing the expression feature information with the preset micro-expression model, and determining the micro-expression in the to-be-identified video according to the comparison result.

In an embodiment, before the step of performing image recognition on the video to be recognized, obtaining a face in the to-be-identified video, and dividing the face according to a preset area, the method further includes:

Extracting environment characteristic information of the to-be-identified video;

Correspondingly, comparing the expression feature information with the preset micro-expression model, specifically:

And comparing the expression feature information and the environment feature information to the preset micro-expression model.

In an embodiment, the step of obtaining the face part in the to-be-identified video specifically includes:

And cutting the to-be-identified video, and retaining a part of the face in the to-be-identified video;

The face portion is segmented to eliminate video segments that do not contain micro-expressions.

In an embodiment, the step of extracting the expression feature information of each preset area in the to-be-identified video includes:

Performing contour recognition on the facial features to obtain contour feature information of the facial features;

Performing a texture analysis on the nasolabial region to obtain texture feature information of the nasolabial region;

Obtaining area feature information of the eyelid region;

The contour feature information, the texture feature information, and the area feature information are respectively used as expression feature information corresponding to the preset region.

In an embodiment, before the acquiring the to-be-identified video, the method further includes:

Performing expression recognition on the sample video to determine a micro-expression in the sample video;

Performing image recognition on the sample video, obtaining a face in the sample video, and dividing a face in the sample video according to a preset area;

Extracting expression feature information of each preset area from the sample video;

Establishing a mapping relationship between the micro-expression and the expression feature information, and storing the mapping relationship to obtain a micro-expression library;

A micro-expression model is established, and the micro-expression model is trained through the mapping relationship to form a preset micro-expression model.

In an embodiment, after the step of performing facial expression recognition on the sample video to determine the micro-expression in the sample video, the method further includes:

Extracting environmental feature information in the sample video;

Correspondingly, the establishing a mapping relationship between the micro-expression and the expression feature information includes:

Establishing a mapping relationship between the micro-expression and the expression feature information and the environment feature information.

In an embodiment, before the step of performing expression recognition on the sample video, the method further includes:

Sorting the sample video according to a type of a person in the sample video, the character type including at least one of each preset age group, gender, and identity type;

Correspondingly, the storing the mapping relationship to obtain a micro-expression library further includes:

The mapping relationship is stored according to the character type to obtain each type of micro-expression library.

In addition, in order to achieve the above object, the present application further provides a micro-expression recognition device, including: a memory, a processor, and a micro-expression recognition stored on the memory and operable on the processor a program that, when executed by the processor, implements the steps of the micro-expression recognition method as described above.

In addition, in order to achieve the above object, the present application further provides a storage medium on which a micro-expression recognition program is stored, and when the micro-expression recognition program is executed by a processor, the micro-expression recognition method as described above is implemented. step.

The present invention performs image recognition on the video to be recognized, obtains a face in the to-be-identified video, and divides the face according to a preset area; and extracts expression feature information of each preset area from the to-be-identified video. And comparing the expression feature information with the preset micro-expression model, and determining the micro-expression in the to-be-identified video according to the comparison result. Since the video to be recognized adopted in the present embodiment is acquired in a natural state, and the expression feature information of each preset area of the face is extracted, the recognition of the micro-expression is more accurate, and the true state of the micro-expression can be better reflected.

DRAWINGS

1 is a schematic structural diagram of a micro-expression recognition device in a hardware operating environment according to an embodiment of the present application;

2 is a schematic flowchart of a first embodiment of a micro-expression recognition method according to the present application;

3 is a schematic flowchart of a second embodiment of a micro-expression recognition method according to the present application;

4 is a schematic flow chart of a third embodiment of a micro-expression recognition method according to 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 structural diagram of a micro-expression recognition device in a hardware operating environment according to an embodiment of the present application.

As shown in FIG. 1, the micro-expression recognition device may include a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to implement connection communication between these components. The user interface 1003 can include a display, and the optional user interface 1003 can also include a standard wired interface, a wireless interface. The network interface 1004 can optionally include a standard wired interface, a wireless interface (such as a WI-FI interface). The memory 1005 may be a high speed RAM memory or a stable memory (non-volatile) Memory), such as disk storage. The memory 1005 can also optionally be a storage device independent of the aforementioned processor 1001.

Those skilled in the art will appreciate that the structure illustrated in FIG. 1 does not constitute a definition of a micro-expression recognition device, and may include more or fewer components than those illustrated, or some components may be combined, or different component arrangements.

As shown in FIG. 1, an operating system, a network communication module, a user interface module, and a micro-expression recognition program may be included in the memory 1005 as a storage medium.

In the micro-expression recognition device shown in FIG. 1 , the network interface 1004 is mainly used to connect to other servers for data communication with the other servers; the user interface 1003 is mainly used for connecting to the user terminal and performing data communication with the user terminal; The micro-expression recognition device calls the micro-expression recognition program stored in the memory 1005 by the processor 1001, and performs the following operations:

Comparing the expression feature information with a preset micro-expression model, and determining a micro-expression in the to-be-identified video according to the comparison result.

Further, the processor 1001 may call the micro-expression recognition program stored in the memory 1005, and further perform the following operations:

Obtaining area feature information of the eyelid region;

Extracting environmental feature information in the sample video;

In this embodiment, the image is recognized by the image to be recognized, the face in the to-be-identified video is obtained, and the face is divided according to the preset area; and the expression features of each preset area are extracted from the to-be-recognized video. And comparing the facial expression feature information with the preset micro-expression model, and determining the micro-expression in the to-be-identified video according to the comparison result. Since the video to be recognized adopted in the present embodiment is acquired in a natural state, and the expression feature information of each preset area of the face is extracted, the recognition of the micro-expression is more accurate, and the true state of the micro-expression can be better reflected.

Based on the above hardware structure, an embodiment of the micro-expression recognition method of the present application is proposed.

Referring to FIG. 2, FIG. 2 is a schematic flowchart of a first embodiment of a micro-expression recognition method according to the present application.

In a first embodiment, the micro-expression recognition method comprises the following steps:

Step S10: performing image recognition on the video to be recognized, obtaining a face in the to-be-identified video, and dividing the face according to a preset area;

It can be understood that, in the micro-expression library used by the general micro-expression recognition method, the micro-expressions in the unnatural state such as suppressed expressions are stored, and the true state of the micro-expressions cannot be fully reflected. To overcome the shortcomings, the implementation In the micro-expression recognition method adopted in the example, the micro-expressions in the natural state are adopted, and the micro-expressions are established by the micro-expressions in the natural state, and the micro-expressions to be recognized are identified by using the established micro-expression library. The more obvious difference is that the micro-expressions used in this embodiment are collected in a natural state, rather than being collected in a suppressed unnatural state.

In order to identify the real state of the micro-expression, the to-be-identified video containing the micro-expression in the natural state is acquired, the expression feature information in the to-be-identified video is extracted, and the micro-expression in the to-be-identified video is identified according to the expression feature information. .

It should be understood that, in order to extract the feature feature information in the to-be-identified video, the feature information of each part of the face will be extracted. In this embodiment, a preset area that can display the micro-expression on the face of the person is selected in advance as the Each part, the preset area includes a facial features area, a nasolabial area, and an eyelid area, and the image to be recognized is image-decomposed into a continuous single-frame image to obtain the to-be-identified The face part in the video is divided into the preset part according to the preset area, so as to facilitate the subsequent extraction of the expression feature information of each preset area.

In order to describe the micro-expression of the to-be-identified video more accurately, before the step S10, the method further includes:

It should be noted that the external environment will also affect the micro-expression. Even with the same facial expression information, different micro-expressions will still be produced due to different environments. For example, a person poses a smile in both environments, but in a bright, soft-colored environment, the smile represents a quiet, comfortable micro-expression, and instead, in a dark, narrow, dirty environment. The smile represents a micro-expression of bitter laughter and self-deprecating. Therefore, the embodiment further extracts the environment feature information and combines with the expression feature information to determine the micro-expression in the to-be-identified video, which is more accurate.

In order to conveniently extract the expression feature information, after the step S10, the method further includes:

And clipping the to-be-identified video according to a position of the face in the to-be-identified video, and retaining a face region;

Cut the clipped video to be recognized, and remove the video clip that does not contain the micro-expression.

It can be understood that the duration of the general micro-expression is 1/25~1/5 seconds, and the length of the pre-acquired video to be recognized is generally long, and it is difficult to extract the ephemeral micro-expression, and the video to be recognized is processed into 1 The duration of ~2 seconds can not damage the micro-expression segments, and it is also convenient to extract the expression feature information in the to-be-identified video. Moreover, in addition to the human face, the to-be-identified video includes other background environments. When extracting the facial expression feature information, the micro-expression is not prominent in the image, which affects the extraction effect. Therefore, after the environment feature information is extracted, the to-be-identified video is subjected to pre-processing including cropping and segmentation, and the to-be-identified video is converted into a micro-expression video of 1 to 2 seconds.

First, the video to be recognized is cropped according to the width of the face, for example, centered on the nose, 1.5 times the length of the face is long, and 1.5 times the width of the face is wide, and a rectangular area is formed, according to the rectangular area The image of the to-be-identified video is cropped to obtain a face video.

Secondly, the face video is segmented, and the video segment that does not contain the micro-expression is removed, and the micro-expression video is obtained.

At this point, the micro-expression video of the face is obtained, which facilitates the subsequent extraction of the expression feature information.

Step S20: extracting expression feature information of each preset area from the to-be-identified video;

It can be understood that the expression feature information refers to a set of data information that can reflect the process of changing the micro-expression, including the duration and degree of change of each preset region of the face. Such as the duration of changes in eyebrows, the degree of change in eye contours, etc.

It should be noted that the human micro-expression is presented by all parts of the face. The change of a single part cannot fully explain the micro-expression of the person. For example, when “happy”, the person will not only raise the corner of the mouth, but the mouth will be lifted and the cheeks will be lifted. Wrinkles, eyelid contraction, and "crow's feet" form at the end of the eye. These parts change together to produce a "happy" micro-expression. The parts affecting the human micro-expression mainly include the facial features area, the nasolabial area, and the eyelid area. Therefore, in the present embodiment, the above-mentioned part is selected as the preset area.

In a specific implementation, the to-be-identified video is clipped and segmented, and converted into a micro-expression video, and extracting the expression feature information in the micro-expression video is more convenient and quick. For each preset area, the expression feature information is extracted, that is, the change duration of each preset area and the degree of change of each preset area are extracted.

Step S30: comparing the expression feature information with a preset micro-expression model, and determining a micro-expression in the to-be-identified video according to the comparison result.

It should be noted that, before performing micro-expression recognition in the to-be-identified video, a preset micro-expression model is established, and expression feature information is input in the preset micro-expression model, and the preset micro-expression model can input The expression feature information is identified, the micro-expression corresponding to the expression feature information is obtained, and the micro-expression is output, that is, the micro-expression in the to-be-identified video is recognized.

Referring to FIG. 3, FIG. 3 is a schematic flowchart of a second embodiment of the micro-expression recognition method according to the present application. Based on the embodiment shown in FIG. 2, a second embodiment of the micro-expression recognition method of the present application is proposed.

In the second embodiment, the step S20 specifically includes:

Step S201: performing contour recognition on the facial features area, and acquiring contour feature information of the facial features area;

It can be understood that the facial features area is a main area that affects a human micro-expression, and the facial features area has a clear outline. By contour recognition of the facial features area, contour feature information of the facial features area can be acquired, and the outline characteristic information is obtained. Includes the duration of the change in the profile of the facial features and the degree of change in the profile. The method for the contour recognition may be an edge detection algorithm, which is not limited in this embodiment.

Step S202: performing texture analysis on the nasolabial region to obtain texture feature information of the nasolabial region;

It can be understood that the nasolabial region is an important region that affects the micro-expression of the human, and the nasolabial region has a texture. By performing texture analysis on the nasolabial region, the texture feature information of the nasolabial region can be obtained. The texture characteristic information includes a duration of change of the nasolabial region and a degree of change of the nasolabial fold. The method of the texture analysis may be a grayscale transformation or a binarization, which is not limited in this embodiment.

Step S203: acquiring area feature information of the eyelid region;

It can be understood that the eyelid region is also an important region affecting the human micro-expression, the eyelid region has a skin near the plane, and the area of the eyelid region can be obtained by calculating the area of the eyelid region in each frame of the video image. Characteristic information, the area characteristic information including a change duration of the eyelid region and a degree of change in the eyelid area.

Step S204: The contour feature information, the texture feature information, and the area feature information are respectively used as the expression feature information corresponding to the preset region.

It should be understood that the contour feature information is used as the expression feature information of the facial features, the texture feature information is used as the expression feature information of the nasolabial region, and the area feature information is used as the eyelid region. The expression feature information is summarized, and the expression feature information of all the preset regions is summarized into the expression feature information corresponding to the to-be-identified video.

In this embodiment, different expression methods are used to extract the facial expression feature information of each preset region, which can better capture the change process of the micro-expression, and subsequently identify the to-be-identified according to the expression feature information. The micro-expressions in the video provide the basis.

Referring to FIG. 4, FIG. 4 is a schematic flowchart of a third embodiment of a micro-expression recognition method according to the present application. Based on the embodiment shown in FIG. 2, a third embodiment of the micro-expression recognition method of the present application is proposed.

In the third embodiment, before the step S10, the method further includes:

Step S001: classify the sample video according to a character type in the sample video, where the character type includes at least one of each preset age group, gender, and identity type;

It can be understood that the embodiment provides a micro-expression recognition method, which is applied to the scenario of establishing the micro-expression library and establishing the preset micro-expression model. Pre-establishing a mapping relationship between the micro-expression and the expression feature information, and saving the mapping relationship to obtain a micro-expression library, wherein the micro-expression and the expression feature information in each set of mapping relationships are acquired according to the same sample video. The sample video adopts a video containing a micro-expression in a natural state, and constructs a mapping relationship between the micro-expression and the expression feature information through the micro-expression contained therein. The unique facial expression feature information corresponding to the unique micro-expression and the micro-expression in the sample video is obtained, and the mapping relationship between the micro-expression and the expression feature information corresponding to the sample video is established.

It should be understood that the sample video is classified according to the type of the character, and the feature extraction of the classified video can finally obtain the micro-expression library of each character type. For example, according to the gender classification, the sample video is first divided into a male sample video and a female sample video according to the gender of the person in the sample video, and then the male sample video and the female sample video are separately extracted, and finally the male micro-expression is obtained. Library and female micro-expression library. Similarly, the sample video is classified according to each preset age group and the identity of the person, and the micro-expression library of each preset age group and the micro-expression library of each identity are obtained.

Step S002: performing expression recognition on the sample video to determine a micro-expression in the sample video;

It can be understood that in order to establish a mapping relationship between the micro-expression and the expression feature information, the expression expression in the sample video is determined by performing expression recognition on the sample video. Before the expression recognition, the six basic expressions of the human being are pre-set as the expression category, so that the recognized expressions belong to the expression category, and the six basic expressions include surprise, disgust, anger, fear, sadness and pleasure. All human expressions can be included in these six basic expression ranges. Of course, it is also possible to subdivide the expression into more types of expressions as the expression category, which is not limited in this embodiment.

Step S003: Extracting environmental feature information in the sample video;

It should be noted that the environment has an influence on the micro-expression, and the micro-expression in the sample video is determined by the environment feature information and the expression feature information to be more accurate.

Step S004: performing image recognition on the sample video, obtaining a face part in the sample video, and dividing a face part in the sample video according to a preset area;

Step S005: extract expression feature information of each preset area from the sample video.

It can be understood that image recognition is performed on the sample video, a face portion in the sample video is obtained, and the face portion is divided according to a preset region, and image recognition is performed on the video to be recognized. And the process of dividing the face part according to the preset area according to the preset part; and the process of extracting the expression feature information of each preset area from the sample video The process of extracting the expression feature information of each preset area in the recognition video is consistent.

Step S006: establishing a mapping relationship between the micro-expression and the expression feature information and environment feature information, and storing the mapping relationship to obtain a micro-expression library;

It should be understood that, after acquiring the micro-expression in the sample video, and the environment feature information and the expression feature information in the sample video, since the micro-expression, the environment feature information, and the expression feature information belong to the same sample video, A mapping relationship between the micro-expression and the expression feature information and the environment feature information may be established. The mapping relationship is stored to obtain a micro-expression library, and the micro-expression library includes a mapping relationship between the micro-expressions of the character types and the expression feature information and the environment feature information.

Step S007: Establish a micro-expression model, and train the micro-expression model through the mapping relationship to form a preset micro-expression model.

It should be noted that the data such as the mapping relationship stored in the micro-expression library is classified according to the character type, and the data stored in each class is scattered and lacks systematicity. By establishing a model, the model is trained to construct a data context. Can complete the finishing of the data. The micro-expression recognition can be conveniently and quickly performed on the to-be-identified video through the trained preset micro-expression model.

It should be understood that, in order to construct a preset micro-expression model, a micro-expression model is pre-established, and the micro-expression model is trained by the mapping relationship to improve the recognition accuracy of the micro-expression model, and the mapping relationship is The obtained known relationship may be used to train the micro-expression model. When the number of mapping relationships is sufficient, and the number of trainings of the micro-expression model is sufficient by using the mapping relationship, the micro-expression model is accurately determined. The rate can reach a certain standard and become the preset micro-expression model.

It can be understood that the specific process of training the micro-expression model by using the mapping relationship to form a preset micro-expression model is: inputting a set of mapping relationships in the micro-expression model, the micro-expression model according to the Deriving the identification result of the sample video by using the environment feature information and the expression feature information in the mapping relationship, and comparing the recognition result with the micro-expression in the mapping relationship to obtain a comparison result;

When the recognition result is consistent with the micro-expression, the output determination result is true, and the micro-expression model connection right is increased, and the next set of mapping relationships is trained;

When the recognition result is inconsistent with the micro-expression, the output determination result is false, the micro-expression model connection right is reduced, and the micro-expression model is trained again through the mapping relationship until the discrimination result of all the mapping relationships is true.

It should be noted that when the number of the mapping relationships is not enough, and the number of trainings of the micro-expression model is insufficient by using the mapping relationship, the recognition accuracy after the training may still not reach the standard and cannot be preset. The micro-expression model only obtains the trial model. Therefore, when the trial model is used to initially perform micro-expression recognition on the trial video, the trial model is secondarily trained by the mapping relationship corresponding to the trial video, so as to realize the recognition accuracy of the trial model. Can reach the standard.

When the trial model is used to identify the micro-expression in the trial video, the steps of the secondary training specifically include:

Performing expression recognition on the trial video to determine a micro-expression in the trial video;

Extracting environmental feature information in the trial video;

Performing image recognition on the trial video, obtaining a face part in the trial video, and dividing the face part according to a preset area;

Extracting expression feature information of each preset area from the trial video;

Inputting the micro-expression, the environmental feature information, and the expression feature information in the trial video to the trial model, and the trial model obtains the recognition result of the trial video according to the environmental feature information and the expression feature information, and Comparing the recognition result with the micro-expression in the trial video to obtain a comparison result;

When the recognition result is consistent with the micro-expression, the output determination result is true, the trial model connection right is increased, and the correspondence relationship between the environmental feature information, the expression feature information, and the micro-expression in the trial video is established. The corresponding relationship is saved in the micro-expression library, thereby expanding the micro-expression library;

When the recognition result is inconsistent with the micro-expression, the output determination result is false, the trial model connection right is reduced, and the trial model is trained by the correspondence relationship to increase the recognition accuracy of the trial model. Get the preset micro-expression model.

In this embodiment, the sample video including the micro-expression in the natural state is obtained, the character type is classified into the sample video, and the environmental feature information and the expression feature information of the sample video are extracted, and the micro-expression and the environmental feature information are established. The mapping relationship between the expression feature information, the micro-expression library and the micro-expression model including the mapping relationship of each preset type are established, and the micro-expression model is trained by the mapping relationship to improve the recognition of the micro-expression model. Accuracy to achieve recognition of micro-expressions by the preset micro-expression model.

In addition, the embodiment of the present application further provides a storage medium, where the micro-expression recognition program is stored, and when the micro-expression recognition program is executed by the processor, the following operations are implemented:

Further, when the micro-expression recognition program is executed by the processor, the following operations are also implemented:

Obtaining area feature information of the eyelid region;

Extracting environmental feature information in the sample video;

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, method, article, or It also includes other elements that are not explicitly listed, or elements that are inherent to such a process, method, item, or system. An element defined by the phrase "comprising a ..." does not exclude the presence of additional equivalent elements in a process, method, article, or system that includes the element, without further limitation.

The serial numbers of the embodiments of the present application are merely for the description, and do not represent the advantages and disadvantages of the embodiments.

Those skilled in the art can clearly understand the above by the description of the above embodiments. The embodiment method can be implemented by means of software plus a necessary general hardware platform, and of course Hardware, but in many cases the former is a better implementation. Based on such understanding, the technical solution of the present application may be in the form of a software product in essence or in part contributing to the prior art. It is now found that the computer software product is stored in a storage medium (such as ROM/RAM, disk, CD), and includes a plurality of instructions for making a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device). And the like) performing 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 micro-expression recognition method, wherein the method comprises the following steps:

Performing image recognition on the recognized video, obtaining a face part in the to-be-identified video, and dividing the face part according to the preset area;

Extracting expression feature information of each preset area from the to-be-identified video;

Comparing the expression feature information with the preset micro-expression model, and determining the micro-expression in the to-be-identified video according to the comparison result.
The method according to claim 1, wherein the method of performing image recognition on the image to be recognized, obtaining a face in the to-be-recognized video, and dividing the face according to a preset area, the method Also includes:

Extracting environment characteristic information of the to-be-identified video;

Correspondingly, comparing the expression feature information with the preset micro-expression model includes:

And comparing the expression feature information and the environment feature information to the preset micro-expression model.
The method of claim 1, wherein the step of obtaining a face portion in the to-be-identified video comprises:

And cutting the to-be-identified video, and retaining a part of the face in the to-be-identified video;

The face portion is segmented to eliminate video segments that do not contain micro-expressions.
The method according to claim 1, wherein the predetermined area includes a facial detail area, a nasolabial area, and an eyelid area; the expression feature information includes a duration of change of each preset area and a degree of change of each preset area .
The method of claim 4, wherein the step of extracting the expression feature information of each preset area in the to-be-identified video comprises:

Performing contour recognition on the facial features to obtain contour feature information of the facial features;

Performing a texture analysis on the nasolabial region to obtain texture feature information of the nasolabial region;

Obtaining area feature information of the eyelid region;

The contour feature information, the texture feature information, and the area feature information are respectively used as expression feature information corresponding to the preset region.
The method of claim 1, wherein before the obtaining the video to be identified, the method further comprises:

Performing expression recognition on the sample video to determine a micro-expression in the sample video;

Performing image recognition on the sample video, obtaining a face in the sample video, and dividing a face in the sample video according to a preset area;

Extracting expression feature information of each preset area from the sample video;

Establishing a mapping relationship between the micro-expression and the expression feature information, and storing the mapping relationship to obtain a micro-expression library;

A micro-expression model is established, and the micro-expression model is trained through the mapping relationship to form a preset micro-expression model.
The method of claim 6, wherein after the step of performing facial expression recognition on the sample video to determine a micro-expression in the sample video, the method further comprises:

Extracting environmental feature information in the sample video;

Correspondingly, the mapping relationship between the micro-expression and the expression feature information is established, including:

Establishing a mapping relationship between the micro-expression and the expression feature information and the environment feature information.
The method of claim 7, wherein before the step of performing expression recognition on the sample video, the method further comprises:

Sorting the sample video according to a type of a person in the sample video, the character type including at least one of each preset age group, gender, and identity type;

Correspondingly, the storing the mapping relationship to obtain a micro-expression library further includes:

The mapping relationship is stored according to the character type to obtain each type of micro-expression library.
A micro-expression recognition device, wherein the micro-expression recognition device comprises: a memory, a processor, and a micro-expression recognition program stored on the memory and operable on the processor, the micro-expression recognition program being The processor implements the following steps when executed:

Performing image recognition on the recognized video, obtaining a face part in the to-be-identified video, and dividing the face part according to the preset area;

Extracting expression feature information of each preset area from the to-be-identified video;

Comparing the expression feature information with the preset micro-expression model, and determining the micro-expression in the to-be-identified video according to the comparison result.
The micro-expression recognition device of claim 9, wherein the micro-expression recognition program is configured to implement the following steps:

Extracting environment characteristic information of the to-be-identified video;

Correspondingly, comparing the expression feature information with the preset micro-expression model includes:

And comparing the expression feature information and the environment feature information to the preset micro-expression model.
The micro-expression recognition device of claim 9, wherein the micro-expression recognition program is configured to implement the following steps:

And cutting the to-be-identified video, and retaining a part of the face in the to-be-identified video;

The face portion is segmented to eliminate video segments that do not contain micro-expressions.
The micro-expression recognition device according to claim 9, wherein the preset area includes a facial detail area, a nasolabial area, and an eyelid area; the expression feature information includes a change duration of each preset area and each preset area The degree of change.
A micro-expression recognition apparatus according to claim 12, wherein said micro-expression recognition program is arranged to implement the following steps:

Performing contour recognition on the facial features to obtain contour feature information of the facial features;

Performing a texture analysis on the nasolabial region to obtain texture feature information of the nasolabial region;

Obtaining area feature information of the eyelid region;

The contour feature information, the texture feature information, and the area feature information are respectively used as expression feature information corresponding to the preset region.
The micro-expression recognition device of claim 9, wherein the micro-expression recognition program is configured to implement the following steps:

Performing expression recognition on the sample video to determine a micro-expression in the sample video;

Performing image recognition on the sample video, obtaining a face in the sample video, and dividing a face in the sample video according to a preset area;

Extracting expression feature information of each preset area from the sample video;

Establishing a mapping relationship between the micro-expression and the expression feature information, and storing the mapping relationship to obtain a micro-expression library;

A micro-expression model is established, and the micro-expression model is trained through the mapping relationship to form a preset micro-expression model.
The micro-expression recognition apparatus according to claim 14, wherein said micro-expression recognition program is set to implement the following steps:

Extracting environmental feature information in the sample video;

Correspondingly, the mapping relationship between the micro-expression and the expression feature information is established, including:

Establishing a mapping relationship between the micro-expression and the expression feature information and the environment feature information.
The micro-expression recognition device according to claim 15, wherein said micro-expression recognition program is set to implement the following steps:

Sorting the sample video according to a type of a person in the sample video, the character type including at least one of each preset age group, gender, and identity type;

Correspondingly, the storing the mapping relationship to obtain a micro-expression library further includes:

The mapping relationship is stored according to the character type to obtain each type of micro-expression library.
A storage medium, wherein the storage medium stores a micro-expression recognition program, and when the micro-expression recognition program is executed by the processor, the following steps are implemented:

Performing image recognition on the recognized video, obtaining a face part in the to-be-identified video, and dividing the face part according to the preset area;

Extracting expression feature information of each preset area from the to-be-identified video;

Comparing the expression feature information with the preset micro-expression model, and determining the micro-expression in the to-be-identified video according to the comparison result.
The storage medium of claim 17, wherein the micro-expression recognition program is configured to implement the following steps:

Extracting environment characteristic information of the to-be-identified video;

Correspondingly, comparing the expression feature information with the preset micro-expression model includes:

And comparing the expression feature information and the environment feature information to the preset micro-expression model.
The storage medium of claim 17, wherein the micro-expression recognition program is configured to implement the following steps:

And cutting the to-be-identified video, and retaining a part of the face in the to-be-identified video;

The face portion is segmented to eliminate video segments that do not contain micro-expressions.
The storage medium according to claim 17, wherein the predetermined area includes a facial detail area, a nasolabial area, and an eyelid area; the expression feature information includes a change duration of each preset area and a change of each preset area degree.