CN109271930A - Micro- expression recognition method, device and storage medium - Google Patents
Micro- expression recognition method, device and storage medium Download PDFInfo
- Publication number
- CN109271930A CN109271930A CN201811075329.9A CN201811075329A CN109271930A CN 109271930 A CN109271930 A CN 109271930A CN 201811075329 A CN201811075329 A CN 201811075329A CN 109271930 A CN109271930 A CN 109271930A
- Authority
- CN
- China
- Prior art keywords
- micro
- image
- segments
- feature
- target facial
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/174—Facial expression recognition
- G06V40/176—Dynamic expression
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
- G06V40/171—Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The present invention provides a kind of micro- expression recognition method, device and storage medium, the methods, comprising: the face characteristic for detecting target facial image gathered in advance obtains at least five human face characteristic points of the target facial image；According to the human face characteristic point of the target facial image and preset image block rule, dicing treatment is carried out to the target facial image, obtains several segments；Micro- expression classification result is obtained by the convolutional neural networks model pre-established according to several described segments.The above method is according to the human face characteristic point obtained and preset image block rule, dicing treatment is carried out to the target facial image, and identification classification is carried out to several obtained segments after stripping and slicing using convolutional neural networks, speed, the precision of micro- Expression Recognition can be effectively improved, to increase substantially the working efficiency of micro- Expression Recognition.
Description
Technical field
The present invention relates to micro- Expression Recognition technical fields more particularly to a kind of micro- expression recognition method, device and storage to be situated between
Matter.
Background technique
Micro- expression is that the mankind attempt the face that is very of short duration, being unable to autonomous control revealed when oppressive or hiding real feelings
Portion's expression.The difference of it and generic expression is that micro- expression duration is very short, only 1/25 second to 1/5 second.Therefore, mostly
Number people is often difficult to be aware of its presence.This facial expression for quickly occurring being not easy to be noticeable is considered and egodefense machine
It is formed with pass, expresses pent-up mood.But the generation of micro- expression and identification psychology with neuromechanism still in research, and
And the frequency that micro- expression occurs is relatively low, ordinary people is not also high to the recognition capability of micro- expression, a workman wants to do his work well must first sharpen his tools,
A set of micro- Expression Recognition system is developed, micro- expression is very important to conducting a research.
Currently, Affectiva company studies " the micro- expression " of people by deep learning algorithm, what is " micro- expression ",
It is it is opposite with expression, expression is exactly a kind of dominant expression, and " micro- expression " be then it is a kind of potential, it is recessive, potential
Expression.With the help of deep learning, the variation of texture and wrinkle and shape by observing all faces is finally summarized
Provide the difference of smile characteristics and phoney even other micro- expressions.In general, the said firm positions 42 passes on face
Then key point tracks difference of these key points between 0.2 second (1 second about 30 frame picture), smiled, greatly with this to identify
Laugh at, finely, the mark such as puzzled.Specific practice is: point on the face being divided into changeability and non-deformable point, is generally made with nose
Geometrical model (at a distance from the point and point) wash with watercolours of other points with nose anchor point is calculated when user's espressiove changes for anchor point
Feature is contaminated, come micro- expression for determining active user is to smile or surprised etc. with this.Above-mentioned mode is classical global calculation
Method, since it is desired that the feature for calculating distance between various combination point just can determine that micro- expression is.Such as: it smiles, is mouth
The linkage of point and the facial muscle point at angle；Such as surprised: being likely to mouth and open point around card and eyes slightly
Minor change.But the speed of the calculating of global calculation method is slow, the precision of micro- Expression Recognition is lower, leads to micro- Expression Recognition
Working efficiency it is low.
Summary of the invention
Based on this, the invention proposes a kind of micro- expression recognition method, device and storage mediums, can effectively improve micro- table
Speed, the precision of feelings identification, to increase substantially the working efficiency of micro- Expression Recognition.
To achieve the above object, on the one hand the embodiment of the present invention provides a kind of micro- expression recognition method, comprising:
The face characteristic for detecting target facial image gathered in advance obtains at least five people of the target facial image
Face characteristic point；
According to the human face characteristic point of the target facial image and preset image block rule, to the target face
Image carries out dicing treatment, obtains several segments；
Micro- expression classification result is obtained by the convolutional neural networks model pre-established according to several described segments.
Preferably, described regular according to the human face characteristic point of the target facial image and preset image block, it is right
The target facial image carries out dicing treatment, obtains several segments, specifically includes:
A plurality of first image segmentation lines are established on the target facial image according to the human face characteristic point；
According to preset image block rule, the cutting of the first image of N item is extracted from a plurality of first image segmentation lines
Line, and using resulting N item the first image segmentation lines of extraction to the target facial image dicing treatment；Wherein, N < 3；
It goes through all over a plurality of first image segmentation lines, obtains several segments altogether.
Preferably, the method also includes:
It is identified in the human face characteristic point not according to several described segments by the convolutional neural networks model
Deformation behaviour point and deformation behaviour point；Wherein, the indeformable characteristic point is in the human face characteristic point not by the volume
The point of neuronal activation, the deformation behaviour point are in the human face characteristic point by the convolutional Neural in product neural network model
The point of neuronal activation in network model；
According to the deformation behaviour point in the human face characteristic point, a plurality of second image is established on the target facial image
Segmentation lines；
According to described image piecemeal rule, the second image of M item segmentation lines are extracted from a plurality of second image segmentation lines,
And using resulting M item the second image segmentation lines of extraction to the target facial image dicing treatment；Wherein, M < 3；
It goes through all over a plurality of second image segmentation lines, obtains several deformation behaviour point segments altogether；
Micro- expression classification is obtained by the convolutional neural networks model according to several described deformation behaviour point segments
Adjustment is as a result, and be updated to micro- expression classification adjustment result for current micro- expression classification result.
Preferably, described several segments according to obtain micro- table by the convolutional neural networks model pre-established
Mutual affection class is as a result, specifically include:
Gray processing processing is carried out to several described segments, obtains several gray processing segments；
Flip horizontal processing is carried out to several described segments, obtains several flip horizontal segments；
Several described segments, several described gray processing segments and several described flip horizontal segments are input to
The convolutional neural networks model carries out convolutional calculation, obtains the corresponding feature vector of the target facial image, and to described
Feature vector carries out PCA dimension-reduction treatment；
According to the feature vector after dimensionality reduction, by the Multilayer Classifier in the convolutional neural networks model, described in acquisition
The corresponding micro- expression classification result of target facial image.
Preferably, described regular according to the human face characteristic point of the target facial image and preset image block, it is right
The target facial image carries out dicing treatment, before obtaining several segments, the method also includes:
It is described that registration process is carried out to the target facial image according to the human face characteristic point.
Preferably, several described segments include: the first segment comprising eyes feature, comprising eyes feature and nose spy
Second segment of sign, the third segment comprising left eye feature and left nose wing feature, the comprising right eye feature and right wing of nose feature
Four segments, the 5th segment comprising nose feature and mouth feature, the 6th segment comprising left nose wing feature and left corners of the mouth feature,
The 7th segment comprising right wing of nose feature and right corners of the mouth feature, the 8th segment comprising looks feature, the comprising mouth feature
Nine segments and the tenth segment comprising full face feature.
On the other hand the embodiment of the present invention provides a kind of micro- expression recognition apparatus, comprising:
Facial features localization module obtains the mesh for detecting the face characteristic of target facial image gathered in advance
Mark at least five human face characteristic points of facial image；
Image slice module, for according to the human face characteristic point of the target facial image and preset image block rule
Then, dicing treatment is carried out to the target facial image, obtains several segments；
Micro- Expression Recognition module, for according to several described segments, by the convolutional neural networks model pre-established,
Obtain micro- expression classification result.
Preferably, described image stripping and slicing module includes:
First image segmentation lines establish unit, for being established on the target facial image according to the human face characteristic point
A plurality of first image segmentation lines；
First cutting unit, for being mentioned from a plurality of first image segmentation lines according to preset image block rule
Take N item the first image segmentation lines, and using extract resulting N item the first image segmentation lines to the target facial image stripping and slicing at
Reason；Wherein, N < 3；
First segment acquiring unit obtains several segments for going through all over a plurality of first image segmentation lines altogether.
On the one hand the embodiment of the present invention provides a kind of micro- expression recognition apparatus, including processor, memory and storage
In the memory and it is configured as the computer program executed by the processor, the processor executes the computer
Such as above-mentioned micro- expression recognition method is realized when program.
On the one hand the embodiment of the present invention provides a kind of computer readable storage medium, the computer readable storage medium
Computer program including storage, wherein control the computer readable storage medium institute in computer program operation
Such as above-mentioned micro- expression recognition method is executed in equipment.
Compared with the prior art, the beneficial effect of the embodiment of the present invention is: micro- expression recognition method, comprising: inspection
The face characteristic for surveying target facial image gathered in advance obtains at least five human face characteristic points of the target facial image；
According to the human face characteristic point of the target facial image and preset image block rule, the target facial image is carried out
Dicing treatment obtains several segments；It is obtained according to several described segments by the convolutional neural networks model pre-established
Obtain micro- expression classification result.The above method is according to the human face characteristic point obtained and preset image block rule, to the mesh
It marks facial image and carries out dicing treatment, and identification point is carried out to several obtained segments after stripping and slicing using convolutional neural networks
Class, can effectively improve speed, the precision of micro- Expression Recognition, to increase substantially the working efficiency of micro- Expression Recognition.
Detailed description of the invention
Fig. 1 is a kind of flow diagram for micro- expression recognition method that the embodiment of the present invention one provides；
Fig. 2 is the stripping and slicing schematic diagram of target facial image；
Fig. 3 is the schematic diagram of several segments obtained after dicing treatment；
A kind of Fig. 4 schematic block diagram of micro- expression recognition apparatus provided by Embodiment 2 of the present invention；
A kind of schematic block diagram for micro- expression recognition apparatus that Fig. 5 embodiment of the present invention three provides.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Embodiment one
Referring to Fig. 1, it is a kind of flow diagram of micro- expression recognition method provided in an embodiment of the present invention.The side
Method can be executed by micro- expression recognition apparatus, specifically includes the following steps:
S100: detecting the face characteristic of target facial image gathered in advance, obtains the target facial image at least
Five human face characteristic points；
In embodiments of the present invention, micro- expression recognition apparatus can be computer, mobile phone, tablet computer, access control equipment, pen
Remember that this computer or server etc. calculate equipment, micro- expression recognition method can be used as that one of functional module is integrated and institute
It states on micro- expression recognition apparatus, is executed by micro- expression recognition apparatus.
In embodiments of the present invention, micro- expression recognition apparatus receives target facial image, it should be noted that this hair
The bright acquisition modes for the target facial image do not do any restrictions, such as can be by micro- expression recognition apparatus certainly
The video camera of band is obtained, and either receives the target from network or other equipment by wired mode or wireless mode
Facial image, micro- expression recognition apparatus carry out the target facial image after receiving the target facial image
Face datection, to obtain at least five human face characteristic points of the target facial image, i.e., the described target facial image includes more
A image pattern, such as the sequential frame image sample extracted from image/video are acquired corresponding each frame image pattern
At least five human face characteristic points, such as extract eyes (2 eyes), nose, the corners of the mouth (2 corners of the mouths).Further, base can be passed through
The target facial image is done symmetrically in the center of two points in two centers and mouth circle, to obtain more face characteristics
Point.Micro- expression recognition apparatus can also directly receive image/video and carry out the pumping of frame image pattern to image/video
It takes, to the target facial image obtained after extraction, comprising: several frames are continuous or the identical image pattern progress in interval is following micro-
Expression Recognition processing.
S200: according to the human face characteristic point of the target facial image and preset image block rule, to the mesh
It marks facial image and carries out dicing treatment, obtain several segments；
In embodiments of the present invention, such as according to the resulting eyes (2 eyes) of extraction, nose, the corners of the mouth (2 corners of the mouths) 5
A human face characteristic point, according to setting image block rule, to eyes (2 eyes), the nose, mouth in the target facial image
Angle (2 corners of the mouths) carries out region combination, such as eyes region, left eye+left nose wing area, right eye+right wing of nose region, mouth, the right side
The wing of nose+right corners of the mouth etc. carries out dicing treatment to the target facial image according to the combination of each region, to obtain several segments.
On the one hand it on the other hand can pass through different people by carrying out cutting to the target facial image with expanding data sample
Face characteristic point combination zone can reduce the selection of key point, improve the precision of micro- Expression Recognition.
S300: micro- expression classification is obtained by the convolutional neural networks model pre-established according to several described segments
As a result.
In embodiments of the present invention, micro- expression recognition apparatus is using several described segments as the convolution pre-established
The input value of neural network model, and the Harr feature vector of 160 dimensions of each segment is calculated, finally to the spy of each segment
Sign vector is combined (the namely splicing of vector), ultimately forms the number of features of 160*Q.And finally using the convolution mind
Classify through the softmax classifier in network model to feature vector, specifically, softmax classifier according to setting in advance
The micro- expression set characteristic type (usually 10 kinds or more, for example, happily, it is painful, sad, startled, angry, angry, feel uncertain, detest
Dislike, is helpless, tired, despise, is self-doubt etc.) feature vector of input is judged, since softmax is multi-categorizer,
The input of same group of data may have probability in multiple classifiers, the corresponding classification of maximum probability is extracted in the present embodiment
The output result of device is as micro- expression classification result.The present invention cuts target facial image by the human face characteristic point extracted
Block can reduce the selection of key point, and the segment extracted in conjunction with convolutional neural networks to stripping and slicing can be mentioned effectively into convolutional calculation
Speed, the precision of high micro- Expression Recognition, to increase substantially the working efficiency of micro- Expression Recognition.
In an alternative embodiment, S200: according to the human face characteristic point of the target facial image and preset
Image block rule carries out dicing treatment to the target facial image, obtains several segments, specifically include:
A plurality of first image segmentation lines are established on the target facial image according to the human face characteristic point；
As shown in Fig. 2, according to resulting eyes (2 eyes), 5 nose, the corners of the mouth (2 corners of the mouths) face characteristics is extracted
Point establishes a plurality of first image segmentation lines, such as: with the first connected image segmentation lines of side point, left nose pterion in left eye angle
L1；First image segmentation lines L made of being connected with bridge of the nose point with prenasale2；It is connected with side point, right wing of nose point in right eye angle and is formed
The first image segmentation lines L3；First image segmentation lines L made of being connected with the top key point of eyes key point4；With eyes
First image segmentation lines L made of the point most on the lower of key point is connected5；With nose most under point, be parallel to the first image and cut
Separated time L4Made of the first image segmentation lines L6。
According to preset image block rule, the cutting of the first image of N item is extracted from a plurality of first image segmentation lines
Line, and using resulting N item the first image segmentation lines of extraction to the target facial image dicing treatment；Wherein, N < 3；
In the present embodiment, 1 in a plurality of first image segmentation lines or 2 the first image segmentation lines can arbitrarily be chosen
To the target facial image dicing treatment.
It goes through all over a plurality of first image segmentation lines, obtains several segments altogether.
In an alternative embodiment, several described segments include: the first segment comprising eyes feature, comprising double
Second segment of eye feature and nose feature, the third segment comprising left eye feature and left nose wing feature, comprising right eye feature with
4th segment of right wing of nose feature, includes left nose wing feature and the left corners of the mouth at the 5th segment comprising nose feature and mouth feature
6th segment of feature, the 7th segment comprising right wing of nose feature and right corners of the mouth feature, the 8th segment comprising looks feature, packet
9th segment of the feature containing mouth and the tenth segment comprising full face feature.
As shown in figure 3, can be understood as pre-set face feature rule of combination according to preset image block rule
(for example, comprising eyes feature, comprising eyes feature and nose feature, comprising left eye feature and left nose wing feature, include right eye spy
Sign and right wing of nose feature, comprising nose feature and mouth feature, include left nose wing feature and left corners of the mouth feature, special comprising the right wing of nose
Sign with right corners of the mouth feature, comprising looks feature, comprising mouth feature and include full face feature), first segment (patch) is
It is cut with the first image segmentation lines L5, includes eyes feature；Second segment be with the first image segmentation lines L6 cutting and
At comprising eyes feature and nose feature；Third segment is with the first image segmentation lines L3 and the first image segmentation lines L6 group
It closes, includes left eye feature and left nose wing feature；4th segment is with the first image segmentation lines L1 and the first image cutting
Line L6 is composed, and includes right wing of nose feature and right corners of the mouth feature；5th segment is that the first image segmentation lines L5 is cut,
Include nose feature and mouth feature；6th segment be the first image segmentation lines L5 combined with the first image segmentation lines L3 and
At comprising left nose wing feature and left corners of the mouth feature；7th segment is the first image segmentation lines L5 and the first image segmentation lines L6
It is composed, includes right wing of nose feature and right corners of the mouth feature；8th segment is that the first image segmentation lines L4 is cut, and includes
Looks feature；9th segment is that the first image segmentation lines L4 is cut, and includes mouth feature；Tenth segment is one complete
Figure, including entire face area.By the way that the target image is carried out stripping and slicing extraction according to above-mentioned rule, using convolutional Neural net
Network can preferably learn the feature to different micro- expressions, such as study which human face characteristic point combination when certain micro- expression occurs
Correspondence moves, and the human face characteristic point that neuron is not activated in convolutional neural networks can be considered indeformable characteristic point (certain
There is no move for key point when the micro- expression of kind occurs), deformation behaviour point (micro- table can be considered by the human face characteristic point of neuronal activation
Key point moves when feelings occur).It is cut according to the organ characteristic of micro- expression setting, for example, the position of mouth (is also wanted
Quantity including point), the combination at canthus and nose, the combination of eyes, the combination there are also forehead and eyes, then to different faces
Characteristic point is combined calculating, and this combination calculation is more it can be found that different micro- expressions influence different devices in which kind of degree
The muscular movement of official can more accurately identify characteristic point and previous frame image sample in present image sample on each segment
Whether the characteristic point of each segment has the sign of movement in this, carries out sentencing for micro- expression according to human face characteristic point combination to realize
It is disconnected.
In an alternative embodiment, the method also includes:
It is identified in the human face characteristic point not according to several described segments by the convolutional neural networks model
Deformation behaviour point and deformation behaviour point；Wherein, the indeformable characteristic point is in the human face characteristic point not by the volume
The point of neuronal activation, the deformation behaviour point are in the human face characteristic point by the convolutional Neural in product neural network model
The point of neuronal activation in network model；
According to the deformation behaviour point in the human face characteristic point, a plurality of second image is established on the target facial image
Segmentation lines；
According to described image piecemeal rule, the second image of M item segmentation lines are extracted from a plurality of second image segmentation lines,
And using resulting M item the second image segmentation lines of extraction to the target facial image dicing treatment；Wherein, M < 3；
In the present embodiment, 1 in a plurality of first image segmentation lines or 2 the first image segmentation lines can arbitrarily be chosen
To the target facial image dicing treatment.
It goes through all over a plurality of second image segmentation lines, obtains several deformation behaviour point segments altogether；
Micro- expression classification is obtained by the convolutional neural networks model according to several described deformation behaviour point segments
Adjustment is as a result, and be updated to micro- expression classification adjustment result for current micro- expression classification result.
It is understood that the target image includes several frames continuously or is spaced identical image pattern, it is current to know
It is clipped in smile expression, looks feature is indeformable characteristic point, other features are deformation behaviour point；Then for next frame target person
Face image, can reject the looks feature in image block rule, repeat above-mentioned image block step to remaining deformation behaviour point
Piecemeal extraction is carried out, it, being capable of guide image piecemeal in turn based on the identification of above-mentioned deformation behaviour point and indeformable characteristic point
With the selection of human face characteristic point, the selection of useless human face characteristic point can be effectively reduced, to be further simplified convolutional calculation mistake
Journey greatlys improve the speed and precision of micro- Expression Recognition.
In an alternative embodiment, described several segments according to, pass through the convolutional Neural net pre-established
Network model obtains micro- expression classification as a result, specifically including:
Gray processing processing is carried out to several described segments, obtains several gray processing segments；
Flip horizontal processing is carried out to several described segments, obtains several flip horizontal segments；
Several described segments, several described gray processing segments and several described flip horizontal segments are input to
The convolutional neural networks model carries out convolutional calculation, obtains the corresponding feature vector of the target facial image, and to described
Feature vector carries out PCA dimension-reduction treatment；
According to the feature vector after dimensionality reduction, by the Multilayer Classifier in the convolutional neural networks model, described in acquisition
The corresponding micro- expression classification result of target facial image.
In the embodiment of the present invention, gray proces are carried out to several described segments for example, by using Weighted Average Algorithm.Due to
The segment extracted from target facial image is color image, is specifically made of multiple pixels, and each pixel
All indicated by tri- values of RGB；Gray proces are carried out to each segment, will not influence the texture feature information of facial expression image, and
And each pixel only needs a gray value that can indicate, substantially increases facial expression image treatment effeciency.Specifically, by following
Gray proces Weighted Average Algorithm formula carries out gray proces to each segment:
F (i, j)=0.30R (i, j)+0.59G (i, j)+0.11B (i, j)
Wherein, i, j represent a pixel in the position of two-dimensional space vector, it may be assumed that the i-th row, jth column.
According to above-mentioned formula, the gray value of each pixel in each segment is calculated, value range is 0-255, makes expression
Black-white-gray state is presented in image.
Gray processing and flip horizontal processing are carried out respectively to several described segments, it can further expanding data sample
This, improves the precision of micro- Expression Recognition.
In embodiments of the present invention, based on 10 segments, after gray processing and flip horizontal processing, 30 are obtained
A segment, micro- expression recognition apparatus using this 30 segments as the input value of the convolutional neural networks model pre-established,
And 160 dimension Harr feature vectors of each segment are calculated, (namely vector finally is combined to the vector of each patch
Splicing), ultimately form the number of features of 160*30.Then PCA dimension-reduction treatment is carried out to this 4800 number of features, is formed
The feature vector of 150 dimensions, and be input to softmax classifier and classify.
PCA dimension-reduction treatment includes the following steps:
The feature vector of the target facial image is converted to the matrix of n*m；
Each row in the matrix is subjected to zero averaging processing；
According to the matrix that zero averaging is handled, covariance matrix is calculated, and calculates the feature vector of the covariance matrix
And its corresponding characteristic value；
From top to bottom by rows according to characteristic value size by the feature vector of the covariance matrix, variation square is obtained
Battle array；
K row forms dimensionality reduction matrix before extracting from the transformation matrices, after obtaining the target facial image PCA dimensionality reduction
Feature vector；Wherein, the numerical value of k is determined according to the compressed error of feature vector of the target facial image.
Specifically, according to formula (1), the numerical value of k is determined；
Wherein, m is the number of all feature vectors in preceding k row；A k is chosen, when the threshold value that error < Φ, Φ are setting
(such as 0.01), it is determined that the dimensionality reduction matrix of the preceding k row composition extracted from the transformation matrices meets dimensionality reduction requirement.
It is a dimension by the corresponding feature vector of the target facial image obtained after Face datection registration process
Higher matrix.Higher dimensional matrix be easy to cause low memory in calculating, is also easy to appear overfitting problem；Therefore it is based on PCA
The modes such as (Principle components analysis) processing function convert the corresponding high dimensional feature vector of human face characteristic point to low by dimensionality reduction
The characteristic of dimension space.For example, choose a K based on the above method, so that error < Φ, then it is considered that this m can be with
Receive, otherwise attempts other.It is converted by PCA dimensionality reduction, the corresponding feature vector of the target facial image by being more than originally
4800 dimensions become 150 dimensions, and subsequent classification problem is become a partition problem in 150 dimension spaces, are keeping main
Information it is complete while greatly simplifie calculating process.
In an alternative embodiment, the human face characteristic point and preset figure according to the target facial image
As piecemeal rule, dicing treatment is carried out to the target facial image, before obtaining several segments, the method also includes:
It is described that registration process is carried out to the target facial image according to the human face characteristic point.
For example, registration process is carried out using open source component OpenCV in the embodiment of the present invention, by the target facial image
In each image pattern Face datection come out after, (such as 160*160 pixel) image for being converted under same scale.Specifically
Ground detects the face in each image pattern of the target facial image using the detectMultiScale algorithm of OpenCV
And frame selects.The key point of face is standardized, i.e. side of the searching face in the point of the leftmost side and the point of top side, as image
Edge, other points are translated on the basis of the two edges；Finally divided by (rightmost side-leftmost side) and (lower side-top side), make one
The standard point of face is evenly distributed in the figure of frame choosing, and it is negative to bring is calculated can be reduced as far as excess pixel point
Load.Affine transformation is finally carried out using the getAffineTransform algorithm of OpenCV, and exports the target person of image alignment
Face image.
Using the target facial image as the input value of the candidate region network model pre-established, with from the candidate
Zone-network model obtains target face area image and at least two human face area images, such as extracts eyes (2
Eyes), nose, the corners of the mouth (2 corners of the mouths) area image, in the candidate region network model, according to the ratio of setting with
And the specification of area, a series of regional frames that the target facial image meets condition are obtained, in the selection of this regional frame
Cheng Zhong, and with convolutional layer come selected characteristic, and inhibit to obtain candidate frame from a series of regional frames by non-maximum value, then lead to
It crosses full articulamentum and carries out candidate frame small parameter perturbations, to obtain target face area image and at least two human face regions
Image can directly be generated the region of suggestion using convolutional neural networks, realize area by the candidate region network model
Domain generates network and the weight of sorter network is shared, substantially increases the performance and speed of monitoring.
Compared with the prior art, a kind of beneficial effect of micro- expression recognition method provided in an embodiment of the present invention is:
(1) present invention is according to the human face characteristic point obtained and preset image block rule, to the target face figure
Identification classification is carried out to several obtained segments after stripping and slicing as carrying out dicing treatment, and using convolutional neural networks, it can
Speed, the precision of micro- Expression Recognition are effectively improved, to increase substantially the working efficiency of micro- Expression Recognition；
(2) present invention is cut according to the organ characteristic of micro- expression setting, for example, the position of mouth (will also include point
Quantity), the combination at canthus and nose, then the combination of eyes, the combination there are also forehead and eyes use convolutional neural networks
Calculating is combined to different faces characteristic point, passes through the different micro- expression shadows in which kind of degree of this combination calculation discovery
The muscular movement for ringing Different Organs, can more accurately identify the characteristic point and upper one in present image sample on each segment
Whether the characteristic point of each segment has the sign of movement in frame image pattern, to realize micro- according to human face characteristic point combination progress
The judgement of expression, micro- Expression Recognition precision are high；
(3) identification of the deformation behaviour point and indeformable characteristic point identified the present invention is based on convolutional neural networks model,
It is capable of the selection of guide image piecemeal and human face characteristic point in turn, the selection of useless human face characteristic point can be effectively reduced, into
One step improves the speed and precision of micro- Expression Recognition.
Embodiment two
Referring to Fig. 4, it is a kind of schematic block diagram of micro- expression recognition apparatus provided in an embodiment of the present invention, described device
Include:
Facial features localization module 1 obtains the mesh for detecting the face characteristic of target facial image gathered in advance
Mark at least five human face characteristic points of facial image；
Image slice module 2, for the human face characteristic point and preset image block according to the target facial image
Rule carries out dicing treatment to the target facial image, obtains several segments；
Micro- Expression Recognition module 3, for passing through the convolutional neural networks mould pre-established according to several described segments
Type obtains micro- expression classification result.
In an alternative embodiment, described image stripping and slicing module 2 includes:
First image segmentation lines establish unit, for being established on the target facial image according to the human face characteristic point
A plurality of first image segmentation lines；
First cutting unit, for being mentioned from a plurality of first image segmentation lines according to preset image block rule
Take N item the first image segmentation lines, and using extract resulting N item the first image segmentation lines to the target facial image stripping and slicing at
Reason；Wherein, N < 3；
First segment acquiring unit obtains several segments for going through all over a plurality of first image segmentation lines altogether.
In an alternative embodiment, several described segments include: the first segment comprising eyes feature, comprising double
Second segment of eye feature and nose feature, the third segment comprising left eye feature and left nose wing feature, comprising right eye feature with
4th segment of right wing of nose feature, includes left nose wing feature and the left corners of the mouth at the 5th segment comprising nose feature and mouth feature
6th segment of feature, the 7th segment comprising right wing of nose feature and right corners of the mouth feature, the 8th segment comprising looks feature, packet
9th segment of the feature containing mouth and the tenth segment comprising full face feature.
In an alternative embodiment, described device further include:
Feature point recognition module, for identifying institute by the convolutional neural networks model according to several described segments
State the indeformable characteristic point and deformation behaviour point in human face characteristic point；Wherein, the indeformable characteristic point is that the face is special
Not by the point of neuronal activation in the convolutional neural networks model in sign point, the deformation behaviour point is the face characteristic
By the point of the neuronal activation in the convolutional neural networks model in point；
Second image segmentation lines establish unit, for according to the deformation behaviour point in the human face characteristic point, in the mesh
A plurality of second image segmentation lines are established on mark facial image；
Second cutting unit, for extracting M from a plurality of second image segmentation lines according to described image piecemeal rule
Item the second image segmentation lines, and using resulting M item the second image segmentation lines of extraction to the target facial image dicing treatment；
Wherein, M < 3；
Second segment acquiring unit obtains several deformation behaviours for going through all over a plurality of second image segmentation lines altogether
Point segment；
Micro- expression classification adjustment unit, for passing through the convolutional Neural according to several described deformation behaviour point segments
Network model obtains micro- expression classification adjustment as a result, and current micro- expression classification result is updated to micro- expression classification tune
Whole result.
In an alternative embodiment, micro- Expression Recognition module 3 includes:
Gray scale processing unit obtains several gray processing segments for carrying out gray processing processing to several described segments；
Flip horizontal processing unit obtains several levels for carrying out flip horizontal processing to several described segments
Overturn segment；
PCA dimension-reduction treatment unit, for will several described segments, several described gray processing segments and described several
A flip horizontal segment is input to the convolutional neural networks model and carries out convolutional calculation, and it is corresponding to obtain the target facial image
Feature vector, and to described eigenvector carry out PCA dimension-reduction treatment；
Micro- expression classification unit, for according to the feature vector after dimensionality reduction, by the convolutional neural networks model
Multilayer Classifier obtains the corresponding micro- expression classification result of the target facial image.
In an alternative embodiment, described device further include:
Registration process module is used for according to the human face characteristic point, described to carry out at alignment to the target facial image
Reason.
Embodiment three
It is the schematic diagram for micro- expression recognition apparatus that third embodiment of the invention provides referring to Fig. 5.Micro- Expression Recognition dress
Setting includes: at least one processor 11, such as CPU, at least one network interface 14 or other users interface 13, memory
15, at least one communication bus 12, communication bus 12 is for realizing the connection communication between these components.Wherein, user interface
13 may include optionally USB interface and other standards interface, wireline interface.Network interface 14 optionally may include Wi-
Fi interface and other wireless interfaces.Memory 15 may include high speed RAM memory, it is also possible to further include non-labile deposit
Reservoir (non-volatilememory), for example, at least a magnetic disk storage.Memory 15 optionally may include at least one
A storage device for being located remotely from aforementioned processor 11.
In some embodiments, memory 15 stores following element, executable modules or data structures, or
Their subset or their superset:
Operating system 151 includes various system programs, for realizing various basic businesses and hardware based of processing
Business；
Program 152.
Specifically, processor 11 executes people described in above-described embodiment for calling the program 152 stored in memory 15
Face matching process, such as step S100 shown in FIG. 1.Alternatively, being realized when the processor execution computer program above-mentioned
The function of each module/unit in each Installation practice, such as facial features localization module.
Illustratively, the computer program can be divided into one or more module/units, one or more
A module/unit is stored in the memory, and is executed by the processor, to complete the present invention.It is one or more
A module/unit can be the series of computation machine program instruction section that can complete specific function, and the instruction segment is for describing institute
State implementation procedure of the computer program in micro- expression recognition apparatus.
Micro- expression recognition apparatus can be the meter such as desktop PC, notebook, palm PC and cloud server
Calculate equipment.Micro- expression recognition apparatus may include, but be not limited only to, processor, memory.Those skilled in the art can manage
Solution, the schematic diagram is only the example of micro- expression recognition apparatus, does not constitute the restriction to micro- expression recognition apparatus, can wrap
It includes than illustrating more or fewer components, perhaps combines certain components or different components, such as micro- Expression Recognition dress
Setting can also include input-output equipment, network access equipment, bus etc..
Alleged processor can be central processing unit (Central Processing Unit, CPU), can also be it
His general processor, digital signal processor (Digital Signal Processor, DSP), specific integrated circuit
(Application Specific Integrated Circuit, ASIC), ready-made programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic,
Discrete hardware components etc..General processor can be microprocessor or the processor is also possible to any conventional processor
Deng the processor is the control centre of micro- expression recognition apparatus, utilizes the entire micro- expression of various interfaces and connection
The various pieces of identification device.
The memory can be used for storing the computer program and/or module, and the processor is by operation or executes
Computer program in the memory and/or module are stored, and calls the data being stored in memory, described in realization
The various functions of micro- expression recognition apparatus.The memory can mainly include storing program area and storage data area, wherein storage
It program area can application program needed for storage program area, at least one function (such as sound-playing function, image player function
Deng) etc.；Storage data area, which can be stored, uses created data (such as audio data, phone directory etc.) etc. according to mobile phone.This
Outside, memory may include high-speed random access memory, can also include nonvolatile memory, such as hard disk, memory, insert
Connect formula hard disk, intelligent memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card, flash memory
Block (Flash Card), at least one disk memory, flush memory device or other volatile solid-state parts.
Wherein, if module/unit that micro- expression recognition apparatus integrates is realized in the form of SFU software functional unit simultaneously
When sold or used as an independent product, it can store in a computer readable storage medium.Based on such reason
Solution, the present invention realize all or part of the process in above-described embodiment method, can also instruct correlation by computer program
Hardware complete, the computer program can be stored in a computer readable storage medium, the computer program is in quilt
When processor executes, it can be achieved that the step of above-mentioned each embodiment of the method.Wherein, the computer program includes computer program
Code, the computer program code can be source code form, object identification code form, executable file or certain intermediate forms
Deng.The computer-readable medium may include: any entity or device, record that can carry the computer program code
Medium, USB flash disk, mobile hard disk, magnetic disk, CD, computer storage, read-only memory (ROM, Read-Only Memory), with
Machine access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium etc..
It should be noted that the content that the computer-readable medium includes can be according to legislation and patent practice in jurisdiction
It is required that carrying out increase and decrease appropriate, such as in certain jurisdictions, do not wrapped according to legislation and patent practice, computer-readable medium
Include electric carrier signal and telecommunication signal.
Example IV
The embodiment of the invention provides a kind of computer readable storage medium, the computer readable storage medium includes depositing
The computer program of storage, wherein equipment where controlling the computer readable storage medium in computer program operation
Execute such as above-mentioned micro- expression recognition method.
It should be noted that the apparatus embodiments described above are merely exemplary, wherein described be used as separation unit
The unit of explanation may or may not be physically separated, and component shown as a unit can be or can also be with
It is not physical unit, it can it is in one place, or may be distributed over multiple network units.It can be according to actual
It needs that some or all of the modules therein is selected to achieve the purpose of the solution of this embodiment.In addition, device provided by the invention
In embodiment attached drawing, the connection relationship between module indicate between them have communication connection, specifically can be implemented as one or
A plurality of communication bus or signal wire.Those of ordinary skill in the art are without creative efforts, it can understand
And implement.
The above is a preferred embodiment of the present invention, it is noted that for those skilled in the art
For, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also considered as
Protection scope of the present invention.
Claims (10)
1. a kind of micro- expression recognition method characterized by comprising
The face characteristic for detecting target facial image gathered in advance, at least five faces for obtaining the target facial image are special
Sign point；
According to the human face characteristic point of the target facial image and preset image block rule, to the target facial image
Dicing treatment is carried out, several segments are obtained；
Micro- expression classification result is obtained by the convolutional neural networks model pre-established according to several described segments.
2. micro- expression recognition method as described in claim 1, which is characterized in that the people according to the target facial image
Face characteristic point and preset image block rule, carry out dicing treatment to the target facial image, obtain several segments,
It specifically includes:
A plurality of first image segmentation lines are established on the target facial image according to the human face characteristic point；
According to preset image block rule, the first image of N item segmentation lines are extracted from a plurality of first image segmentation lines, and
Using resulting N item the first image segmentation lines of extraction to the target facial image dicing treatment；Wherein, N < 3；
It goes through all over a plurality of first image segmentation lines, obtains several segments altogether.
3. micro- expression recognition method as claimed in claim 2, which is characterized in that the method also includes:
It is identified indeformable in the human face characteristic point according to several described segments by the convolutional neural networks model
Characteristic point and deformation behaviour point；Wherein, the indeformable characteristic point is in the human face characteristic point not by the convolution mind
Point through neuronal activation in network model, the deformation behaviour point are in the human face characteristic point by the convolutional neural networks
The point of neuronal activation in model；
According to the deformation behaviour point in the human face characteristic point, a plurality of second image cutting is established on the target facial image
Line；
According to described image piecemeal rule, the second image of M item segmentation lines are extracted from a plurality of second image segmentation lines, and adopt
With resulting M item the second image segmentation lines of extraction to the target facial image dicing treatment；Wherein, M < 3；
It goes through all over a plurality of second image segmentation lines, obtains several deformation behaviour point segments altogether；
Micro- expression classification adjustment is obtained by the convolutional neural networks model according to several described deformation behaviour point segments
As a result, and current micro- expression classification result is updated to micro- expression classification adjustment result.
4. micro- expression recognition method as described in claim 1, which is characterized in that described several segments according to pass through
The convolutional neural networks model pre-established obtains micro- expression classification as a result, specifically including:
Gray processing processing is carried out to several described segments, obtains several gray processing segments；
Flip horizontal processing is carried out to several described segments, obtains several flip horizontal segments；
Several described segments, several described gray processing segments and several described flip horizontal segments are input to described
Convolutional neural networks model carries out convolutional calculation, obtains the corresponding feature vector of the target facial image, and to the feature
Vector carries out PCA dimension-reduction treatment；
The target is obtained by the Multilayer Classifier in the convolutional neural networks model according to the feature vector after dimensionality reduction
The corresponding micro- expression classification result of facial image.
5. micro- expression recognition method as described in claim 1, which is characterized in that the people according to the target facial image
Face characteristic point and preset image block rule, carry out dicing treatment to the target facial image, obtain several segments
Before, the method also includes:
It is described that registration process is carried out to the target facial image according to the human face characteristic point.
6. micro- expression recognition method as claimed in claim 1 or 2, which is characterized in that several described segments include: comprising double
First segment of eye feature, includes left eye feature and left nose wing feature at the second segment comprising eyes feature and nose feature
Third segment, the 4th segment comprising right eye feature and right wing of nose feature, the 5th segment comprising nose feature and mouth feature,
The 6th segment comprising left nose wing feature and left corners of the mouth feature, the 7th segment comprising right wing of nose feature and right corners of the mouth feature, packet
The 8th segment, the 9th segment comprising mouth feature and the tenth segment comprising full face feature of the feature containing looks.
7. a kind of micro- expression recognition apparatus characterized by comprising
Facial features localization module obtains the target person for detecting the face characteristic of target facial image gathered in advance
At least five human face characteristic points of face image；
Image slice module, for regular according to the human face characteristic point of the target facial image and preset image block,
Dicing treatment is carried out to the target facial image, obtains several segments；
Micro- Expression Recognition module, for being obtained according to several described segments by the convolutional neural networks model pre-established
Micro- expression classification result.
8. micro- expression recognition apparatus as claimed in claim 7, which is characterized in that described image stripping and slicing module includes:
First image segmentation lines establish unit, a plurality of for being established on the target facial image according to the human face characteristic point
First image segmentation lines；
First cutting unit, for extracting N item from a plurality of first image segmentation lines according to preset image block rule
First image segmentation lines, and using resulting N item the first image segmentation lines of extraction to the target facial image dicing treatment；Its
In, N < 3；
First segment acquiring unit obtains several segments for going through all over a plurality of first image segmentation lines altogether.
9. a kind of micro- expression recognition apparatus, which is characterized in that including processor, memory and storage in the memory and
It is configured as the computer program executed by the processor, the processor realizes such as right when executing the computer program
It is required that micro- expression recognition method described in any one of 1 to 6.
10. a kind of computer readable storage medium, which is characterized in that the computer readable storage medium includes the calculating of storage
Machine program, wherein equipment where controlling the computer readable storage medium in computer program operation is executed as weighed
Benefit require any one of 1 to 6 described in micro- expression recognition method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811075329.9A CN109271930B (en) | 2018-09-14 | 2018-09-14 | Micro-expression recognition method, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811075329.9A CN109271930B (en) | 2018-09-14 | 2018-09-14 | Micro-expression recognition method, device and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109271930A true CN109271930A (en) | 2019-01-25 |
| CN109271930B CN109271930B (en) | 2020-11-13 |
Family
ID=65189116
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811075329.9A Active CN109271930B (en) | 2018-09-14 | 2018-09-14 | Micro-expression recognition method, device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109271930B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110046955A (en) * | 2019-03-12 | 2019-07-23 | 平安科技（深圳）有限公司 | Marketing method, device, computer equipment and storage medium based on recognition of face |
| CN110147729A (en) * | 2019-04-16 | 2019-08-20 | 深圳壹账通智能科技有限公司 | User emotion recognition methods, device, computer equipment and storage medium |
| CN111178151A (en) * | 2019-12-09 | 2020-05-19 | 量子云未来（北京）信息科技有限公司 | Method and device for realizing human face micro-expression change recognition based on AI technology |
| CN112329663A (en) * | 2020-11-10 | 2021-02-05 | 西南大学 | Micro-expression time detection method and device based on face image sequence |
| CN112511748A (en) * | 2020-11-30 | 2021-03-16 | 努比亚技术有限公司 | Lens target intensified display method and device, mobile terminal and storage medium |
| CN112668384A (en) * | 2020-08-07 | 2021-04-16 | 深圳市唯特视科技有限公司 | Knowledge graph construction method and system, electronic equipment and storage medium |
| WO2021082045A1 (en) * | 2019-10-29 | 2021-05-06 | 平安科技（深圳）有限公司 | Smile expression detection method and apparatus, and computer device and storage medium |
| CN113792572A (en) * | 2021-06-17 | 2021-12-14 | 重庆邮电大学 | Facial expression recognition method based on local representation |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8848068B2 (en) * | 2012-05-08 | 2014-09-30 | Oulun Yliopisto | Automated recognition algorithm for detecting facial expressions |
| CN106295566A (en) * | 2016-08-10 | 2017-01-04 | 北京小米移动软件有限公司 | Facial expression recognizing method and device |
| CN106570474A (en) * | 2016-10-27 | 2017-04-19 | 南京邮电大学 | Micro expression recognition method based on 3D convolution neural network |
| CN106599800A (en) * | 2016-11-25 | 2017-04-26 | 哈尔滨工程大学 | Face micro-expression recognition method based on deep learning |
| US20180092549A1 (en) * | 2015-06-14 | 2018-04-05 | Facense Ltd. | Detecting physiological responses based on multispectral data from head-mounted cameras |
-
2018
- 2018-09-14 CN CN201811075329.9A patent/CN109271930B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8848068B2 (en) * | 2012-05-08 | 2014-09-30 | Oulun Yliopisto | Automated recognition algorithm for detecting facial expressions |
| US20180092549A1 (en) * | 2015-06-14 | 2018-04-05 | Facense Ltd. | Detecting physiological responses based on multispectral data from head-mounted cameras |
| CN106295566A (en) * | 2016-08-10 | 2017-01-04 | 北京小米移动软件有限公司 | Facial expression recognizing method and device |
| CN106570474A (en) * | 2016-10-27 | 2017-04-19 | 南京邮电大学 | Micro expression recognition method based on 3D convolution neural network |
| CN106599800A (en) * | 2016-11-25 | 2017-04-26 | 哈尔滨工程大学 | Face micro-expression recognition method based on deep learning |
Non-Patent Citations (2)
| Title |
|---|
| M. A. TAKALKAR ET AL.: "Image Based Facial Micro-Expression Recognition Using Deep Learning on Small Datasets", 《2017 INTERNATIONAL CONFERENCE ON DIGITAL IMAGE COMPUTING: TECHNIQUES AND APPLICATIONS (DICTA)》 * |
| 王勋: "基于组合特征的人脸表情识别算法研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110046955A (en) * | 2019-03-12 | 2019-07-23 | 平安科技（深圳）有限公司 | Marketing method, device, computer equipment and storage medium based on recognition of face |
| CN110147729A (en) * | 2019-04-16 | 2019-08-20 | 深圳壹账通智能科技有限公司 | User emotion recognition methods, device, computer equipment and storage medium |
| WO2021082045A1 (en) * | 2019-10-29 | 2021-05-06 | 平安科技（深圳）有限公司 | Smile expression detection method and apparatus, and computer device and storage medium |
| CN111178151A (en) * | 2019-12-09 | 2020-05-19 | 量子云未来（北京）信息科技有限公司 | Method and device for realizing human face micro-expression change recognition based on AI technology |
| CN112668384A (en) * | 2020-08-07 | 2021-04-16 | 深圳市唯特视科技有限公司 | Knowledge graph construction method and system, electronic equipment and storage medium |
| CN112668384B (en) * | 2020-08-07 | 2024-05-31 | 深圳市唯特视科技有限公司 | Knowledge graph construction method, system, electronic equipment and storage medium |
| CN112329663A (en) * | 2020-11-10 | 2021-02-05 | 西南大学 | Micro-expression time detection method and device based on face image sequence |
| CN112329663B (en) * | 2020-11-10 | 2023-04-07 | 西南大学 | Micro-expression time detection method and device based on face image sequence |
| CN112511748A (en) * | 2020-11-30 | 2021-03-16 | 努比亚技术有限公司 | Lens target intensified display method and device, mobile terminal and storage medium |
| CN113792572A (en) * | 2021-06-17 | 2021-12-14 | 重庆邮电大学 | Facial expression recognition method based on local representation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109271930B (en) | 2020-11-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109271930A (en) | Micro- expression recognition method, device and storage medium | |
| US10579862B2 (en) | Method, device, and computer readable storage medium for detecting feature points in an image | |
| CN107679448B (en) | Eyeball action-analysing method, device and storage medium | |
| Bougourzi et al. | Deep learning based face beauty prediction via dynamic robust losses and ensemble regression | |
| CN107633204A (en) | Face occlusion detection method, apparatus and storage medium | |
| CN111898547A (en) | Training method, device and equipment of face recognition model and storage medium | |
| CN109145871A (en) | Psychology and behavior recognition methods, device and storage medium | |
| CN109859305A (en) | Three-dimensional face modeling, recognition methods and device based on multi-angle two-dimension human face | |
| CN109829448A (en) | Face identification method, device and storage medium | |
| CN109978918A (en) | A kind of trajectory track method, apparatus and storage medium | |
| CN109271884A (en) | Face character recognition methods, device, terminal device and storage medium | |
| CN108765278A (en) | A kind of image processing method, mobile terminal and computer readable storage medium | |
| KR20230107415A (en) | Method for identifying an object within an image and mobile device for executing the method | |
| CN111274916A (en) | Face recognition method and face recognition device | |
| CN109145837A (en) | Face emotion identification method, device, terminal device and storage medium | |
| CN112215180A (en) | Living body detection method and device | |
| CN108198130B (en) | Image processing method, image processing device, storage medium and electronic equipment | |
| CN110414428A (en) | A method of generating face character information identification model | |
| CN107633205A (en) | lip motion analysis method, device and storage medium | |
| CN108875540A (en) | Image processing method, device and system and storage medium | |
| CN104484658A (en) | Face gender recognition method and device based on multi-channel convolution neural network | |
| CN110163111A (en) | Method, apparatus of calling out the numbers, electronic equipment and storage medium based on recognition of face | |
| CN108960167A (en) | Hair style recognition methods, device, computer readable storage medium and computer equipment | |
| CN112084917A (en) | Living body detection method and device | |
| CN106897659A (en) | The recognition methods of blink motion and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |