CN107798281A - A kind of human face in-vivo detection method and device based on LBP features - Google Patents
A kind of human face in-vivo detection method and device based on LBP features Download PDFInfo
- Publication number
- CN107798281A CN107798281A CN201610808240.3A CN201610808240A CN107798281A CN 107798281 A CN107798281 A CN 107798281A CN 201610808240 A CN201610808240 A CN 201610808240A CN 107798281 A CN107798281 A CN 107798281A
- Authority
- CN
- China
- Prior art keywords
- lbp
- grader
- feature
- facial image
- features
- 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
- G06V40/161—Detection; Localisation; Normalisation
-
- 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
- G06V40/40—Spoof detection, e.g. liveness detection
- G06V40/45—Detection of the body part being alive
Landscapes
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a kind of human face in-vivo detection method and device based on LBP features, methods described includes：The near-infrared facial image and visible ray facial image of the tested face of collection；Near-infrared facial image and visible ray facial image are pre-processed respectively；The first LBP features of pretreated near-infrared facial image and the 2nd LBP features of pretreated visible ray facial image are extracted respectively；By the first LBP features and the 2nd LBP features input the first grader of cascade respectively and the second grader is classified, and judge whether tested face is live body according to classification results.The human face in-vivo detection method of the present invention does not need the motion of user, action to coordinate, and user experience is good；Also, due to the LBP features being extracted under two kinds of different acquisition patterns of visible ray and near-infrared, consider the two and carry out multi-cascade classification, improve the accuracy rate and robustness of In vivo detection.
Description
Technical field
The present invention relates to technical field of biometric identification, more particularly to a kind of face In vivo detection side based on LBP features
Method and device.
Background technology
Recognition of face, it is a kind of biological identification technology that the facial feature information based on people carries out identification.With shooting
Machine or camera collection image or video flowing containing face, and automatic detect and track face in the picture, and then to detection
The face that arrives carries out a series of correlation techniques of face, generally also referred to as Identification of Images, face recognition.The life of modern people and
In work, it can effectively strengthen safety and privacy using face recognition technology.But there is also a problem, such as profit in reality
Face recognition device may be cheated with the false information such as photo, mask or video so that disabled user passes through recognition of face
This road security perimeter, constitutes a threat to safety and privacy.
Therefore, before recognition of face, first carry out In vivo detection, can effectively prevent using printing photo, mobile phone or
The deceptive information such as photo and video, face mask in the mobile terminals such as Pad are by recognition of face, so as to avoid security breaches.
Existing In vivo detection technology mainly has：It is based on texture, based on light stream and based on interactive In vivo detection etc..
Wherein, the different three-dimensional structures such as live body true man, photo, screen are obtained by establishing light stream field model based on the method for light stream
Light stream feature in motion, carries out In vivo detection, shortcoming is to need user to have certain motion.Examined based on interactive live body
Survey, it is necessary to which detected object completes the action of system instruction, such as rotary head, the action such as blink, open one's mouth, but these specific actions can
To be deceived by recorded video or other modes so as to In vivo detection system of out-tricking.Both above inspection mode living since it is desired that
User movement or action coordinate, not friendly enough, and user experience is poor.Biopsy method based on texture, only extract visible ray
The textural characteristics of lower user's picture, the feature of extraction are limited, and accuracy is not high, robustness is bad.
The content of the invention
The invention provides a kind of human face in-vivo detection method and device based on LBP features, is examined with solving existing live body
The problem of survey technology user experience is poor, accuracy and robustness are relatively low.
In order to solve the above problems, the invention provides a kind of human face in-vivo detection method based on LBP features, including：
The near-infrared facial image and visible ray facial image of the tested face of collection；
Near-infrared facial image and visible ray facial image are pre-processed respectively；
The first LBP features of pretreated near-infrared facial image and pretreated visible ray face figure are extracted respectively
2nd LBP features of picture；
By the first LBP features and the 2nd LBP features input the first grader of cascade respectively and the second grader enters
Row classification, judge whether tested face is live body according to classification results.
As one for example, described input the first of cascade respectively by the first LBP features and the 2nd LBP features
Grader and the second grader are classified, including：First LBP features are inputted after the first grader classified, then by the
Two LBP features input the second grader and classified；Or after the 2nd LBP features the second grader of input is classified, then
First LBP features are inputted into the first grader to be classified.
As one for example, first grader includes：The first sub-classifier and the second sub-classifier of cascade；
First sub-classifier is the grader trained by the near-infrared face sample image of live body, photo；Second subclassification
Device is the grader trained by the near-infrared face sample image of live body, mask.
As one for example, second grader includes：The 3rd sub-classifier and the 4th sub-classifier of cascade；
3rd sub-classifier is the grader trained by the visible ray face sample image of live body, photo；4th subclassification
Device is the grader trained by the visible ray face sample image of live body, mask.
As one for example, the first LBP features include：59 dimensionsWhat feature, 59 were tieed upIt is special
What sign, 243 were tieed upOne or several kinds of combinations in feature；Wherein,Be characterized as neighborhood territory pixel point number for 8,
The LBP histogram features that the uniform pattern LBP algorithms that the radius of neighbourhood is 1 obtain；It is characterized as that neighborhood territory pixel point number is
8th, the LBP histogram features that the uniform pattern LBP algorithms that the radius of neighbourhood is 2 obtain；It is characterized as neighborhood territory pixel point number
The LBP histogram features that the uniform pattern LBP algorithms for being 2 for the 16, radius of neighbourhood obtain.
As one for example, the 2nd LBP features include：59 dimensionsWhat feature, 59 were tieed upIt is special
What sign, 243 were tieed upOne or several kinds of combinations in feature；Wherein,It is characterized as that neighborhood territory pixel point number is
8th, the LBP histogram features that the uniform pattern LBP algorithms that the radius of neighbourhood is 1 obtain；It is characterized as neighborhood territory pixel point number
The LBP histogram features that the uniform pattern LBP algorithms for being 2 for the 8, radius of neighbourhood obtain；It is characterized as neighborhood territory pixel point
The LBP histogram features that the uniform pattern LBP algorithms that number is 16, the radius of neighbourhood is 2 obtain.
Accordingly, present invention also offers a kind of face living body detection device based on LBP features, including：
Image acquisition units, for gathering the near-infrared facial image and visible ray facial image of tested face；
Graphics processing unit, for being pre-processed respectively to near-infrared facial image and visible ray facial image；
Feature extraction unit, for the first LBP features for extracting pretreated near-infrared facial image respectively and pre- place
2nd LBP features of the visible ray facial image after reason；
Classification judging unit, for the first LBP features and the 2nd LBP features to be inputted to the first classification of cascade respectively
Device and the second grader are classified, and judge whether tested face is live body according to classification results.
As one for example, the classification judging unit is used to the first LBP features inputting the progress of the first grader
After classification, then the 2nd LBP features are inputted into the second grader and classified；Or the 2nd LBP features are inputted into the second grader
After being classified, then the first LBP features are inputted into the first grader and classified.
As one for example, first grader includes：The first sub-classifier and the second sub-classifier of cascade；
First sub-classifier is the grader trained by the near-infrared face sample image of live body, photo；Second subclassification
Device is the grader trained by the near-infrared face sample image of live body, mask.
As one for example, second grader includes：The 3rd sub-classifier and the 4th sub-classifier of cascade；
3rd sub-classifier is the grader trained by the visible ray face sample image of live body, photo；4th subclassification
Device is the grader trained by the visible ray face sample image of live body, mask.
As one for example, the first LBP features include：59 dimensionsWhat feature, 59 were tieed upIt is special
What sign, 243 were tieed upOne or several kinds of combinations in feature；Wherein,Be characterized as neighborhood territory pixel point number for 8,
The LBP histogram features that the uniform pattern LBP algorithms that the radius of neighbourhood is 1 obtain；It is characterized as that neighborhood territory pixel point number is
8th, the LBP histogram features that the uniform pattern LBP algorithms that the radius of neighbourhood is 2 obtain；It is characterized as neighborhood territory pixel point number
The LBP histogram features that the uniform pattern LBP algorithms for being 2 for the 16, radius of neighbourhood obtain.
As one for example, the 2nd LBP features include：59 dimensionsWhat feature, 59 were tieed upIt is special
What sign, 243 were tieed upOne or several kinds of combinations in feature；Wherein,Be characterized as neighborhood territory pixel point number for 8,
The LBP histogram features that the uniform pattern LBP algorithms that the radius of neighbourhood is 1 obtain；It is characterized as that neighborhood territory pixel point number is
8th, the LBP histogram features that the uniform pattern LBP algorithms that the radius of neighbourhood is 2 obtain；It is characterized as neighborhood territory pixel point number
The LBP histogram features that the uniform pattern LBP algorithms for being 2 for the 16, radius of neighbourhood obtain.
Compared with prior art, the present invention has advantages below：
In visible ray facial image of the present invention except gathering tested face, pretreated visible ray facial image is extracted
The 2nd LBP features, also acquire the near-infrared facial image of tested face, and be extracted pretreated near-infrared face figure
First LBP features of picture, the first LBP features and the 2nd LBP features are inputted to the first grader and the second classification of cascade respectively
Device is classified, and to realize In vivo detection, the process does not need the motion of user, action to coordinate, and user experience is good.Also,
The textural characteristics being extracted under different photoenvironments, i.e. the first LBP features of pretreated near-infrared facial image and pre- place
2nd LBP features of the visible ray facial image after reason, consider the two and carry out multi-cascade classification, the texture for participating in classification is special
Sign is more abundant, comprehensive, enhances the accuracy rate and robustness of In vivo detection.
In addition, the first grader of the present invention can include：The first sub-classifier and the second sub-classifier of cascade；Pass through
First grader is made into further classification to split, using the near-infrared facial image of live body as positive sample, photo, mask it is near red
Outer facial image is input in different sub-classifiers, for the image of near-infrared collection, energy respectively as different negative samples
It is enough that more accurately to live body, photo and mask, these different types of non-living bodies divide, improve In vivo detection accuracy,
Robustness.
Accordingly, the second grader of the invention can include：The 3rd sub-classifier and the 4th sub-classifier of cascade.It is logical
Cross and make further classification to the second grader and split, using the visible ray facial image of live body as positive sample, photo, mask can
See that light facial image respectively as different negative samples, is input in different sub-classifiers, for the image of visible light collection,
Can more accurately to live body, photo and mask, these different types of non-living bodies divide, can equally improve live body inspection
Accuracy, the robustness of survey.
Brief description of the drawings
Fig. 1 is a kind of flow chart of the human face in-vivo detection method embodiment one based on LBP features of the present invention；
Fig. 2 is a kind of flow chart of the human face in-vivo detection method embodiment two based on LBP features of the present invention；
Fig. 3 is a kind of structural representation of the face living body detection device embodiment based on LBP features of the present invention.
Embodiment
In order to facilitate the understanding of the purposes, features and advantages of the present invention, it is below in conjunction with the accompanying drawings and specific real
Applying mode, the present invention is further detailed explanation.
Reference picture 1, show a kind of flow of the human face in-vivo detection method embodiment one based on LBP features of the present invention
Figure, methods described include：
The near-infrared facial image and visible ray facial image of step 101, the tested face of collection；
The tested face is probably living body faces, it is also possible to non-living body face, for example, photo, video, mask etc.,
The tested image of face in both modes is gathered using bimodulus image acquisition units, the bimodulus image acquisition units can wrap
Include：Near-infrared light compensating lamp, near-infrared camera and visible image capturing head.Wherein, it is seen that light video camera head has visible filter,
It can filter out the near infrared light of some wave bands, through visible ray, and visible ray facial image is gathered using visible image capturing head；
Near-infrared camera has near infrared filter, and it can filter out visible ray, the near infrared light through some wave bands, using near
Infrared camera gathers near-infrared facial image；Near-infrared light compensating lamp carries out near-infrared light filling, may be provided at near-infrared camera
Around.
The gatherer process of image can be simultaneously collection or acquisition time, for example, first gather near-infrared face figure
Picture, then visible ray facial image is gathered, or, near-infrared facial image is first gathered, then gather visible ray facial image.It can manage
Solution, should be directed to same tested face and gather its corresponding near-infrared facial image and visible ray facial image.
Step 102, near-infrared facial image and visible ray facial image are pre-processed respectively；
The original facial image of collection tends not to directly use due to being limited and random disturbances by various conditions,
Image preprocessing must be carried out to it, the process of feature extraction can be applied to.For facial image, it is pre-processed
Process mainly includes greyscale transformation and normalization.Near-infrared facial image and visible ray facial image are after gray scale conversion, figure
The pixel of picture has the gray value between 0~255, then, by the image normalization after gray scale conversion to the big of fixation
It is small, for example, the size of the pixels of 64 pixel * 64 can be normalized to.In addition, according to different demands, preprocessing process can be with
Light compensation, geometric correction, filtering and sharpening including facial image etc..
Step 103, the first LBP features for extracting pretreated near-infrared facial image respectively and pretreated visible
2nd LBP features of light facial image；
Because illumination reflectivity is different, the texture information appearance that them can be caused to be imaged is poor for living body faces, photo, mask etc.
Different, the embodiment of the present invention is classified using this species diversity to carry out live body and non-living body, specifically, the texture information utilized is LBP
(Local Binary Pattern, local binary patterns) feature, it can characterize texture information well and have certain light
According to consistency.LBP is a kind of operator for being used for describing image local textural characteristics, and it has rotational invariance and gray scale consistency
Etc. it is notable the advantages of, for texture feature extraction.
The process for extracting LBP features is as follows：Firstly, for each pixel, the gray value of itself and adjacent pixel is entered
Row compares, if the gray value of adjacent pixel is more than the gray value of center pixel, the position of the neighbor pixel is marked as 1,
Otherwise it is 0.So, adjacent pixel can form a binary number, that is, obtain the LBP values of central pixel point；Then, it is sharp
The LBP histogram features corresponding to the acquisition of LBP values；Using histogram as a characteristic vector, the LBP of view picture figure has also just been obtained
Feature.
This step extracts the LBP features (that is, the first LBP features) of pretreated near-infrared facial image and pre- place respectively
The LBP features (that is, the 2nd LBP features) of visible ray facial image after reason, it is to be understood that the first LBP features and second
The extraction process of LBP features can be that progress simultaneously or timesharing are carried out.
Step 104, the first grader that the first LBP features and the 2nd LBP features are inputted to cascade respectively and the second classification
Device is classified, and judges whether tested face is live body according to classification results.
In this step, the first grader is used to classify to the first LBP features, and the second grader is used for second
LBP features are classified.The first grader, the second grader in cascade classifier are all SVM (Support Vector
Machine, SVMs) grader.In machine learning field, SVM is a learning model for having supervision, be commonly used into
Row mode identification, classification and regression analysis.First grader, the second grader are preset to be trained by sample image
Obtained grader, it can in advance gather and obtain the LBP characteristics of substantial amounts of living body faces sample and non-living body face sample
According to training function svmtrain (linear nuclear parameter) to train grader with Matlab SVM.Specifically：
For the first grader, sample image is near-infrared face sample image, the near-infrared face of non-living body of live body
Sample image, wherein, the photo of non-living body including various situations (for example tile, fold, bending, deducting eyes face etc.) and each
Kind mask.Obtain the LBP features of the near-infrared face sample image of more parts of live bodies up to ten thousand, the near-infrared people of more parts of non-living bodies up to ten thousand
The LBP features of face sample image, it is input to the first grader and carries out SVM training, and labeled bracketing result, wherein, mark is lived
Body face is+1, and non-living body face is -1.
For the second grader, sample image is visible ray face sample image, the visible ray face of non-living body of live body
Sample image, wherein, the photo of non-living body including various situations (for example tile, fold, bending, deducting eyes face etc.) and each
Kind mask.Obtain the LBP features of the visible ray face sample image of more parts of live bodies up to ten thousand, the visible ray people of more parts of non-living bodies up to ten thousand
The LBP features of face sample image, it is input to the second grader and carries out SVM training, and labeled bracketing result, wherein, mark is lived
Body face is+1, and non-living body face is -1.
In the inventive method embodiment one, the visible ray facial image except gathering tested face, after extraction pretreatment
Visible ray facial image the 2nd LBP features, also acquire the near-infrared facial image of tested face, and be extracted pretreatment
First LBP features of near-infrared facial image afterwards, the first LBP features and the 2nd LBP features are inputted the first of cascade respectively
Grader and the second grader are classified, and to realize In vivo detection, the process does not need the motion of user, action to coordinate, and uses
Family Experience Degree is good.Also, employ the first LBP features of pretreated near-infrared facial image, pretreated visible ray
2nd LBP features of facial image, consider the two and carry out multi-cascade classification, due to being extracted two kinds of visible ray and near-infrared
LBP features under different acquisition pattern, the textural characteristics for participating in classification more enrich, comprehensively, enhance the accurate of In vivo detection
Rate and robustness.
It is understood that the waterfall sequence of the first grader and the second grader can be adjusted successively, and accordingly, first
The order that LBP features and the 2nd LBP features are input to corresponding grader can also be adjusted successively.
As one for example, in this step, the first LBP features can be inputted into the first grader and be classified
Afterwards, then the 2nd LBP features are inputted into the second grader to be classified, if all by the way that is, double classification result is+1, then
Judge tested face as live body；Otherwise, then judge tested face as non-living body.Specifically, the first LBP features are inputted first point
Class device is classified first, when classification results are non-living body first, then detection of end process；When classification results are live body first,
The 2nd LBP features are inputted into the second grader again and carry out secondary classification, when secondary classification result is non-living body, then detection of end mistake
Journey；It is when secondary classification result is live body, then final to judge tested face as live body.
As another for example, in this step, the 2nd LBP features can be inputted into the second grader and be classified
Afterwards, then the first LBP features are inputted into the first grader to be classified, if all by the way that is, double classification result is+1, then
Judge tested face as live body；Otherwise, then judge tested face as non-living body.
As one for example, first grader includes：The first sub-classifier and the second sub-classifier of cascade；
Wherein, first sub-classifier is the grader trained by the near-infrared face sample image of live body, photo；Second son
Grader is the grader trained by the near-infrared face sample image of live body, mask.It is further by making to the first grader
Classification is split, and using the near-infrared facial image of live body as positive sample, photo, the near-infrared facial image of mask be not respectively as
Same negative sample, is input in different sub-classifiers, can be more accurately to live body, photograph for the image of near-infrared collection
Piece and mask these different types of non-living bodies are divided, and improve accuracy, the robustness of In vivo detection.It is appreciated that
It is that the waterfall sequence of the first sub-classifier and the second sub-classifier can be adjusted successively, for example, can be defeated by the first LBP features
Enter the first sub-classifier to be classified, and then the first LBP features are inputted into the second sub-classifier and classified；Or by
One LBP features input the second sub-classifier and classified, and then the first LBP features are inputted into the first sub-classifier and divided
Class.
As another for example, second grader includes：The 3rd sub-classifier and the 4th subclassification of cascade
Device；Wherein, the 3rd sub-classifier is the grader trained by the visible ray face sample image of live body, photo；Described
Four sub-classifiers are the grader trained by the visible ray face sample image of live body, mask.By the second grader is made into
The classification of one step is split, and is made the visible ray facial image of live body as positive sample, photo, the visible ray facial image of mask respectively
For different negative samples, it is input in different sub-classifiers, can be more accurately to work for the image of visible light collection
These different types of non-living bodies of body, photo and mask are divided, and improve accuracy, the robustness of In vivo detection.It can manage
The waterfall sequence of solution, the 3rd sub-classifier and the 4th sub-classifier can be adjusted successively, for example, can be special by the 2nd LBP
Sign the 3rd sub-classifier of input is classified, and then the 2nd LBP features are inputted into the 4th sub-classifier and classified；Or
2nd LBP features are inputted into the 4th sub-classifier to be classified, and then the 2nd LBP features are inputted into the 3rd sub-classifier and entered
Row classification.
As one for example, the first LBP features include：59 dimensionsWhat feature, 59 were tieed upIt is special
What sign, 243 were tieed upOne or several kinds of combinations in feature.
As one for example, the 2nd LBP features include：59 dimensionsWhat feature, 59 were tieed upIt is special
What sign, 243 were tieed upOne or several kinds of combinations in feature.
Wherein,It is characterized as that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 8, the radius of neighbourhood is 1 obtain
LBP histogram features；It is characterized as that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 8, the radius of neighbourhood is 2 obtain
The LBP histogram features arrived；It is characterized as that the uniform pattern LBP that neighborhood territory pixel point number is 16, the radius of neighbourhood is 2 is calculated
The LBP histogram features that method obtains.
Below, withIt is specifically described exemplified by feature：For each pixel in pretreated facial image,
Compare it and the gray value size that the radius of neighbourhood is 8 field pixels around 1, if during the gray value of field pixel is more than
The gray value of imago vegetarian refreshments, then the position of the field pixel be marked as 1, be otherwise 0.In the direction of the clock by 8 numerals
It is linked to be the 2 system numbers of one 8, shares 256 kinds；Obtain 82 system numbers are joined end to end, formed a ring, by its 0 to 1
And 1 to 0 change frequency is classified as one kind, referred to as Uniform patterns (that is, uniform pattern or mould of equal value no more than 2 times
Formula), wherein, u2 is to represent the Uniform patterns that change frequency is no more than 2 times.Through counting, totally 58 kinds of uniform pattern LBP operators,
Non-homogeneous LBP operators are calculated as a kind, then with one 59 dimension vector description neighborhood territory pixel point number be 8, the radius of neighbourhood be 1 it is equal
The LBP histogram features that even pattern LBP algorithms obtain.
Feature,The acquisition process of feature withThe acquisition process of feature is similar, and belongs to existing
There is the content of technology, here is omitted.
As one for example, it is described 59 dimensionWhat feature, 59 were tieed upWhat feature, 243 were tieed upFeature can be the feature obtained for the statistics of whole near-infrared facial image and visible ray facial image.For example,
The first LBP features can be 59 dimensionsFeature；Can also be 59 dimensionsFeature and 59 dimensions
The combination of feature, form the characteristic vector of one 118 dimension；It can also be 59 dimensionsWhat feature, 59 were tieed upIt is special
What sign, 243 were tieed upThe combination of feature, form the characteristic vector of one 361 dimension.For example, the 2nd LBP features can be with
It is 59 dimensionsFeature；Can also be 59 dimensionsFeature and 59 dimensionsThe combination of feature, form one
The characteristic vector of 118 dimensions；It can also be 59 dimensionsWhat feature, 59 were tieed upWhat feature, 243 were tieed upFeature
Combination, formed one 361 dimension characteristic vector.
As another for example, whole near-infrared facial image and visible ray facial image can also be respectively divided
Into different zonules (cell), for example, 9 pieces of zonules are divided into, 59 dimensionWhat feature, 59 were tieed up
What feature, 243 were tieed upAt least one of feature is the feature obtained respectively for different small area statistics, then
The statistic histogram of obtained each zonule is attached as the characteristic vector of a multidimensional.It is it should be noted that each
Individual zonule can mutually be connected with border, and each zonule can also have overlapped place.For example, the first LBP
Feature includes：59 dimensionsWhat feature, 59 were tieed upWhat feature, 243 were tieed upThe combination of feature, wherein, will
Near-infrared facial image is divided into 9 zonules according to preset rules, and its corresponding 59 dimension is obtained for each zonuleFeature, then, and whole infrared face imageThe dimension of feature is tieed up for 59*9=531, the first LBP features
Dimension is：531+59+243=833 is tieed up.
Reference picture 2, show a kind of flow of the human face in-vivo detection method embodiment two based on LBP features of the present invention
Figure, methods described include：
Step 201, the near-infrared image and visible images for gathering measured object；
Image of this step using bimodulus collecting unit collection measured object in both modes, the bimodulus collecting unit can
With including：Near-infrared light compensating lamp, near-infrared camera and visible image capturing head.
It whether there is face in step 202, detection near-infrared image；If so, then perform step 203；If it is not, then judge do not have
It is non-living body face to have face or tested face, and is terminated；
In this method embodiment two, first, the face in near-infrared image is detected using near-infrared Face datection algorithm,
Near-infrared image rather than visible images are first detected, are because of some photos, such as ink-jet photographic, and mobile phone, Pad or notebook
It will not be imaged Deng the video in medium in near-infrared environment, so can first filter part non-living body, improve the standard of algorithm
True property.
Step 203, the corresponding position of detection visible images whether there is face；If so, then perform step 204；If
It is no, then judge no face, and terminate；
Position relationship and near-infrared according to the angle of visual field of visible image capturing head and near-infrared camera, therebetween
The face location detected in image, the regional extent of face in visible images can be extrapolated, in this regional extent
Detect whether face be present, relative to the detection of face in whole visible images, due to reducing detection range, speed can be more
It is quicker, while also ensure that In vivo detection is directed to same tested face, avoid in image mutual when there are multiple faces
Situation about obscuring.
Step 204, near-infrared image and visible images are pre-processed respectively；
By step 202 and 203 detection, confirm to be respectively provided with face in the image of collection, then near-infrared referred to herein
Image is near-infrared facial image, and alleged visible images are visible ray facial image.The pretreatment includes：Gray scale
Conversion and normalized.
First LBP features of step 205, the pretreated near-infrared image of extraction；
Step 206, by the first LBP features input the first sub-classifier classified, determined whether according to classification results
Live body；If so, then perform step 207；If it is not, being then judged as non-living body, and terminate；
First LBP features of pretreated near-infrared facial image are input to the first sub-classifier, first son
Grader is the grader trained by the near-infrared face sample image of live body, photo, is classified by the first order, in judged result
During to be, tentative tested face is live body and is further processed；When judged result is no, can identify as non-live
The Part of photos taken of body.
Step 207, by the first LBP features input the second sub-classifier classified, determined whether according to classification results
Live body；If so, then perform step 208；If it is not, being then judged as non-living body, and terminate；
First LBP features of pretreated near-infrared facial image are input to the second sub-classifier, second son
Grader is the grader trained by the near-infrared face sample image of live body, mask, is classified by the second level, in judged result
During to be, tentative tested face is live body and is further processed；When judged result is no, can identify as non-live
The mask segment of body.
2nd LBP features of step 208, the pretreated visible images of extraction；
Step 209, by the 2nd LBP features input the 3rd sub-classifier classified, determined whether according to classification results
Live body；If so, then perform step 210；If it is not, being then judged as non-living body, and terminate；
2nd LBP features of pretreated visible ray facial image are input to the 3rd sub-classifier, the 3rd son
Grader is the grader trained by the visible ray face sample image of live body, photo, is classified by the third level, in judged result
During to be, tentative tested face is live body and is further processed；When judged result is no, step 206 can not known
Not Chu Lai photo detection come out, as non-living body.
Step 210, by the 2nd LBP features input the 4th sub-classifier classified, determined whether according to classification results
Live body；If so, then judging tested face as live body, and terminate；If it is not, being then judged as non-living body, and terminate.
2nd LBP features of pretreated visible ray facial image are input to the 4th sub-classifier, the 4th son
Grader is the grader trained by the visible ray face sample image of live body, mask, is classified by the fourth stage, in judged result
During to be, final testing result is obtained, it is live body to be tested face；When judged result is no, step 207 can not identified
Mask out detects, as non-living body.
, can be by near-infrared facial image a part by level Four grader in the inventive method embodiment two
Photo, mask are identified, another part photo, mask are identified by visible ray facial image, and then are judged tested
Whether face is live body.
As one for example, the step 206, the sequencing of step 207 can adjust, the step 209, step
Rapid 210 sequencing can also adjust.As another for example, step 208~step 210 can be first carried out, then hold
Row step 205~step 207.As another for example, before step 208 can be mentioned to step 206, i.e. carry respectively
Take the first LBP features of pretreated near-infrared facial image and the 2nd LBP spies of pretreated visible ray facial image
Sign and then execution step 206, step 207, step 209 and step 210.
Reference picture 3, show that a kind of structure of the face living body detection device embodiment two based on LBP features of the present invention is shown
It is intended to, described device 300 includes：
Image acquisition units 301, for gathering the near-infrared facial image and visible ray facial image of tested face；
Graphics processing unit 302, for being pre-processed respectively to near-infrared facial image and visible ray facial image；
Feature extraction unit 303, for extracting the first LBP features of pretreated near-infrared facial image and pre- respectively
2nd LBP features of the visible ray facial image after processing；
Classification judging unit 304, for the first LBP features and the 2nd LBP features to be inputted into the first of cascade respectively
Grader and the second grader are classified, and judge whether tested face is live body according to classification results.
In apparatus of the present invention embodiment, tested face is gathered in both of which using the image acquisition units 301 of bimodulus
Under image, the image acquisition units 301 of the bimodulus can include：Near-infrared light compensating lamp, near-infrared camera and visible ray
Camera.Wherein, it is seen that light video camera head has visible filter, and it can filter out the near infrared light of some wave bands, pass through
Visible ray, visible ray facial image is gathered using visible image capturing head；Near-infrared camera has near infrared filter, and it can be with
Visible ray, the near infrared light through some wave bands are filtered out, near-infrared facial image is gathered using near-infrared camera；Near-infrared
Light compensating lamp carries out near-infrared light filling, may be provided at around near-infrared camera.
The preprocessing process of graphics processing unit 302 mainly includes greyscale transformation and normalization.Classify the of judging unit 304
One grader is used to classify to the first LBP features, and the second grader is used to classify to the 2nd LBP features.Described
One grader, the second grader are the preset graders for training to obtain by sample image, can in advance gather and obtain a large amount of
Living body faces sample and non-living body face sample LBP characteristics, train function svmtrain (lines with Matlab SVM
Property nuclear parameter) train grader.Specifically：
For the first grader, sample image is near-infrared face sample image, the near-infrared face of non-living body of live body
Sample image, wherein, the photo of non-living body including various situations (for example tile, fold, bending, deducting eyes face etc.) and each
Kind mask.For the second grader, sample image is visible ray face sample image, the visible ray face sample of non-living body of live body
This image, wherein, the photo of non-living body including various situations (for example tile, fold, bending, deducting eyes face etc.) and various
Mask.
In apparatus of the present invention embodiment, the visible ray facial image except gathering tested face, extract pretreated
2nd LBP features of visible ray facial image, also acquire the near-infrared facial image of tested face, and after being extracted pretreatment
Near-infrared facial image the first LBP features, the first LBP features and the 2nd LBP features are inputted first point of cascade respectively
Class device and the second grader are classified, and to realize In vivo detection, the process does not need the motion of user, action to coordinate, user
Experience Degree is good.Also, employ the first LBP features of pretreated near-infrared facial image, pretreated visible ray people
2nd LBP features of face image, consider the two and carry out multi-cascade classification, due to being extracted two kinds of visible ray and near-infrared not
With the LBP features under drainage pattern, the textural characteristics for participating in classification more enrich, comprehensively, enhance the accuracy rate of In vivo detection
And robustness.
As one for example, the classification judging unit 304 is used to enter the first LBP features the first grader of input
After row classification, then the 2nd LBP features are inputted into the second grader and classified；Or the 2nd LBP features input second is classified
After device is classified, then the first LBP features are inputted into the first grader and classified.
As one for example, first grader includes：The first sub-classifier and the second sub-classifier of cascade；
First sub-classifier is the grader trained by the near-infrared face sample image of live body, photo；Second subclassification
Device is the grader trained by the near-infrared face sample image of live body, mask.
As one for example, second grader includes：The 3rd sub-classifier and the 4th sub-classifier of cascade；
3rd sub-classifier is the grader trained by the visible ray face sample image of live body, photo；4th subclassification
Device is the grader trained by the visible ray face sample image of live body, mask.
As one for example, the first LBP features include：59 dimensionsWhat feature, 59 were tieed upIt is special
What sign, 243 were tieed upOne or several kinds of combinations in feature；Wherein,Be characterized as neighborhood territory pixel point number for 8,
The LBP histogram features that the uniform pattern LBP algorithms that the radius of neighbourhood is 1 obtain；It is characterized as that neighborhood territory pixel point number is
8th, the LBP histogram features that the uniform pattern LBP algorithms that the radius of neighbourhood is 2 obtain；It is characterized as neighborhood territory pixel point
The LBP histogram features that the uniform pattern LBP algorithms that number is 16, the radius of neighbourhood is 2 obtain.
As one for example, the 2nd LBP features include：59 dimensionsWhat feature, 59 were tieed upIt is special
What sign, 243 were tieed upOne or several kinds of combinations in feature；Wherein,It is characterized as that neighborhood territory pixel point number is
8th, the LBP histogram features that the uniform pattern LBP algorithms that the radius of neighbourhood is 1 obtain；It is characterized as neighborhood territory pixel point number
The LBP histogram features that the uniform pattern LBP algorithms for being 2 for the 8, radius of neighbourhood obtain；It is characterized as neighborhood territory pixel point
The LBP histogram features that the uniform pattern LBP algorithms that number is 16, the radius of neighbourhood is 2 obtain.
Each embodiment in this specification is described by the way of progressive, what each embodiment stressed be and its
The difference of his embodiment, between each embodiment identical similar part mutually referring to.For device embodiment
Speech, because it is substantially similar to embodiment of the method, so description is fairly simple, referring to the part of embodiment of the method in place of correlation
Explanation.
Above to a kind of human face in-vivo detection method and device based on LBP features provided by the present invention, carry out in detail
Thin to introduce, specific case used herein is set forth to the principle and embodiment of the present invention, and above example is said
It is bright to be only intended to help the method and its core concept for understanding the present invention；Meanwhile for those of ordinary skill in the art, foundation
The thought of the present invention, there will be changes in specific embodiments and applications, in summary, this specification content is not
It is interpreted as limitation of the present invention.
Claims (12)
- A kind of 1. human face in-vivo detection method based on LBP features, it is characterised in that including：The near-infrared facial image and visible ray facial image of the tested face of collection；Near-infrared facial image and visible ray facial image are pre-processed respectively；The first LBP features of pretreated near-infrared facial image and pretreated visible ray facial image are extracted respectively 2nd LBP features；By the first LBP features and the 2nd LBP features input the first grader of cascade respectively and the second grader is divided Class, judge whether tested face is live body according to classification results.
- 2. the method as described in claim 1, it is characterised in that described to distinguish the first LBP features and the 2nd LBP features The first grader and the second grader for inputting cascade are classified, including：After first LBP features the first grader of input is classified, then the 2nd LBP features are inputted into the second grader and divided Class；Or input the 2nd LBP features after the second grader classified, then the first LBP features are inputted into the first grader Classified.
- 3. method as claimed in claim 2, it is characterised in thatFirst grader includes：The first sub-classifier and the second sub-classifier of cascade；First sub-classifier is the grader trained by the near-infrared face sample image of live body, photo；Second sub-classifier is the grader trained by the near-infrared face sample image of live body, mask.
- 4. method as claimed in claim 2, it is characterised in thatSecond grader includes：The 3rd sub-classifier and the 4th sub-classifier of cascade；3rd sub-classifier is the grader trained by the visible ray face sample image of live body, photo；4th sub-classifier is the grader trained by the visible ray face sample image of live body, mask.
- 5. the method as described in any one of Claims 1-4, it is characterised in that the first LBP features include：59 dimensionsWhat feature, 59 were tieed upWhat feature, 243 were tieed upOne or several kinds of combinations in feature； Wherein,It is characterized as the LBP histograms that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 8, the radius of neighbourhood is 1 obtain Feature；It is characterized as the LBP histograms that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 8, the radius of neighbourhood is 2 obtain Feature；It is characterized as the LBP Nogatas that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 16, the radius of neighbourhood is 2 obtain Figure feature.
- 6. the method as described in any one of Claims 1-4, it is characterised in that the 2nd LBP features include：59 dimensionsWhat feature, 59 were tieed upWhat feature, 243 were tieed upOne or several kinds of groups in feature Close；Wherein,It is characterized as the LBP histograms that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 8, the radius of neighbourhood is 1 obtain Feature；It is characterized as the LBP histograms that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 8, the radius of neighbourhood is 2 obtain Feature；It is characterized as the LBP Nogatas that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 16, the radius of neighbourhood is 2 obtain Figure feature.
- A kind of 7. face living body detection device based on LBP features, it is characterised in that including：Image acquisition units, for gathering the near-infrared facial image and visible ray facial image of tested face；Graphics processing unit, for being pre-processed respectively to near-infrared facial image and visible ray facial image；Feature extraction unit, after the first LBP features and the pretreatment for extracting pretreated near-infrared facial image respectively Visible ray facial image the 2nd LBP features；Classify judging unit, for the first LBP features and the 2nd LBP features are inputted respectively cascade the first grader and Second grader is classified, and judges whether tested face is live body according to classification results.
- 8. device as claimed in claim 7, it is characterised in thatThe classification judging unit is used to input the first LBP features after the first grader classified, then by the 2nd LBP features The second grader is inputted to be classified；Or input the 2nd LBP features after the second grader classified, then by the first LBP Feature inputs the first grader and classified.
- 9. device as claimed in claim 8, it is characterised in thatFirst grader includes：The first sub-classifier and the second sub-classifier of cascade；First sub-classifier is the grader trained by the near-infrared face sample image of live body, photo；Second sub-classifier is the grader trained by the near-infrared face sample image of live body, mask.
- 10. device as claimed in claim 8, it is characterised in thatSecond grader includes：The 3rd sub-classifier and the 4th sub-classifier of cascade；3rd sub-classifier is the grader trained by the visible ray face sample image of live body, photo；4th sub-classifier is the grader trained by the visible ray face sample image of live body, mask.
- 11. the device as described in any one of claim 7 to 10, it is characterised in that the first LBP features include：59 dimensionsWhat feature, 59 were tieed upWhat feature, 243 were tieed upOne or several kinds of combinations in feature； Wherein,It is characterized as the LBP histograms that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 8, the radius of neighbourhood is 1 obtain Feature；It is characterized as the LBP histograms that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 8, the radius of neighbourhood is 2 obtain Feature；It is characterized as the LBP Nogatas that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 16, the radius of neighbourhood is 2 obtain Figure feature.
- 12. the device as described in any one of claim 7 to 10, it is characterised in that the 2nd LBP features include：59 dimensionsWhat feature, 59 were tieed upWhat feature, 243 were tieed upOne or several kinds of groups in feature Close；Wherein,It is characterized as the LBP histograms that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 8, the radius of neighbourhood is 1 obtain Feature；It is characterized as the LBP histograms that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 8, the radius of neighbourhood is 2 obtain Feature；It is characterized as the LBP Nogatas that the uniform pattern LBP algorithms that neighborhood territory pixel point number is 16, the radius of neighbourhood is 2 obtain Figure feature.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610808240.3A CN107798281B (en) | 2016-09-07 | 2016-09-07 | Face living body detection method and device based on LBP (local binary pattern) characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610808240.3A CN107798281B (en) | 2016-09-07 | 2016-09-07 | Face living body detection method and device based on LBP (local binary pattern) characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107798281A true CN107798281A (en) | 2018-03-13 |
| CN107798281B CN107798281B (en) | 2021-10-08 |
Family
ID=61530858
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610808240.3A Active CN107798281B (en) | 2016-09-07 | 2016-09-07 | Face living body detection method and device based on LBP (local binary pattern) characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107798281B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108776786A (en) * | 2018-06-04 | 2018-11-09 | 北京京东金融科技控股有限公司 | Method and apparatus for generating user's truth identification model |
| CN108830229A (en) * | 2018-06-20 | 2018-11-16 | 哈尔滨理工大学 | The vivo identification method of Face datection is combined under a kind of frame based on caffe |
| CN108875546A (en) * | 2018-04-13 | 2018-11-23 | 北京旷视科技有限公司 | Face auth method, system and storage medium |
| CN108921041A (en) * | 2018-06-06 | 2018-11-30 | 深圳神目信息技术有限公司 | A kind of biopsy method and device based on RGB and IR binocular camera |
| CN109697417A (en) * | 2018-12-14 | 2019-04-30 | 江阴弘远新能源科技有限公司 | A kind of production management system for pitch-controlled system cabinet |
| CN110008878A (en) * | 2019-03-27 | 2019-07-12 | 中控智慧科技股份有限公司 | A kind of anti-false method of Face datection and the face identification device for having anti-false function |
| CN110008820A (en) * | 2019-01-30 | 2019-07-12 | 广东世纪晟科技有限公司 | A kind of silence biopsy method |
| CN110059607A (en) * | 2019-04-11 | 2019-07-26 | 深圳市华付信息技术有限公司 | Living body multiple detection method, device, computer equipment and storage medium |
| CN110443102A (en) * | 2018-05-04 | 2019-11-12 | 北京眼神科技有限公司 | Living body faces detection method and device |
| CN110580454A (en) * | 2019-08-21 | 2019-12-17 | 北京的卢深视科技有限公司 | Living body detection method and device |
| WO2020077866A1 (en) * | 2018-10-17 | 2020-04-23 | 平安科技（深圳）有限公司 | Moire-based image recognition method and apparatus, and device and storage medium |
| CN111879724A (en) * | 2020-08-05 | 2020-11-03 | 中国工程物理研究院流体物理研究所 | Human skin mask identification method and system based on near infrared spectrum imaging |
| CN112395929A (en) * | 2019-08-19 | 2021-02-23 | 扬州盛世云信息科技有限公司 | Face living body detection method based on infrared image LBP histogram characteristics |
| CN112651268A (en) * | 2019-10-11 | 2021-04-13 | 北京眼神智能科技有限公司 | Method and device for eliminating black and white photos in biopsy, and electronic equipment |
| CN117953591A (en) * | 2024-03-27 | 2024-04-30 | 中国人民解放军空军军医大学 | Intelligent limb rehabilitation assisting method and device |
| CN112651268B (en) * | 2019-10-11 | 2024-05-28 | 北京眼神智能科技有限公司 | Method and device for eliminating black-and-white photo in living body detection and electronic equipment |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101404060A (en) * | 2008-11-10 | 2009-04-08 | 北京航空航天大学 | Human face recognition method based on visible light and near-infrared Gabor information amalgamation |
| CN102163288A (en) * | 2011-04-06 | 2011-08-24 | 北京中星微电子有限公司 | Eyeglass detection method and device |
| CN102867188A (en) * | 2012-07-26 | 2013-01-09 | 中国科学院自动化研究所 | Method for detecting seat state in meeting place based on cascade structure |
| CN104680141A (en) * | 2015-02-13 | 2015-06-03 | 华中师范大学 | Motion unit layering-based facial expression recognition method and system |
| CN105069448A (en) * | 2015-09-29 | 2015-11-18 | 厦门中控生物识别信息技术有限公司 | True and false face identification method and device |
| CN105320950A (en) * | 2015-11-23 | 2016-02-10 | 天津大学 | A video human face living body detection method |
| CN105718868A (en) * | 2016-01-18 | 2016-06-29 | 中国科学院计算技术研究所 | Face detection system and method for multi-pose faces |
| CN105787437A (en) * | 2016-02-03 | 2016-07-20 | 东南大学 | Vehicle brand type identification method based on cascading integrated classifier |
-
2016
- 2016-09-07 CN CN201610808240.3A patent/CN107798281B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101404060A (en) * | 2008-11-10 | 2009-04-08 | 北京航空航天大学 | Human face recognition method based on visible light and near-infrared Gabor information amalgamation |
| CN102163288A (en) * | 2011-04-06 | 2011-08-24 | 北京中星微电子有限公司 | Eyeglass detection method and device |
| CN102867188A (en) * | 2012-07-26 | 2013-01-09 | 中国科学院自动化研究所 | Method for detecting seat state in meeting place based on cascade structure |
| CN104680141A (en) * | 2015-02-13 | 2015-06-03 | 华中师范大学 | Motion unit layering-based facial expression recognition method and system |
| CN105069448A (en) * | 2015-09-29 | 2015-11-18 | 厦门中控生物识别信息技术有限公司 | True and false face identification method and device |
| CN105320950A (en) * | 2015-11-23 | 2016-02-10 | 天津大学 | A video human face living body detection method |
| CN105718868A (en) * | 2016-01-18 | 2016-06-29 | 中国科学院计算技术研究所 | Face detection system and method for multi-pose faces |
| CN105787437A (en) * | 2016-02-03 | 2016-07-20 | 东南大学 | Vehicle brand type identification method based on cascading integrated classifier |
Non-Patent Citations (1)
| Title |
|---|
| JUKKA MÄÄTTÄ 等: ""Face Spoofing Detection From Single Images Using Micro-Texture Analysis"", 《2011 INTERNATIONAL JOINT CONFERENCE ON BIOMETRICS》 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108875546A (en) * | 2018-04-13 | 2018-11-23 | 北京旷视科技有限公司 | Face auth method, system and storage medium |
| CN110443102A (en) * | 2018-05-04 | 2019-11-12 | 北京眼神科技有限公司 | Living body faces detection method and device |
| CN110443102B (en) * | 2018-05-04 | 2022-05-24 | 北京眼神科技有限公司 | Living body face detection method and device |
| CN108776786A (en) * | 2018-06-04 | 2018-11-09 | 北京京东金融科技控股有限公司 | Method and apparatus for generating user's truth identification model |
| CN108921041A (en) * | 2018-06-06 | 2018-11-30 | 深圳神目信息技术有限公司 | A kind of biopsy method and device based on RGB and IR binocular camera |
| CN108830229A (en) * | 2018-06-20 | 2018-11-16 | 哈尔滨理工大学 | The vivo identification method of Face datection is combined under a kind of frame based on caffe |
| WO2020077866A1 (en) * | 2018-10-17 | 2020-04-23 | 平安科技（深圳）有限公司 | Moire-based image recognition method and apparatus, and device and storage medium |
| CN109697417A (en) * | 2018-12-14 | 2019-04-30 | 江阴弘远新能源科技有限公司 | A kind of production management system for pitch-controlled system cabinet |
| CN110008820A (en) * | 2019-01-30 | 2019-07-12 | 广东世纪晟科技有限公司 | A kind of silence biopsy method |
| CN110008878B (en) * | 2019-03-27 | 2021-07-30 | 熵基科技股份有限公司 | Anti-fake method for face detection and face recognition device with anti-fake function |
| CN110008878A (en) * | 2019-03-27 | 2019-07-12 | 中控智慧科技股份有限公司 | A kind of anti-false method of Face datection and the face identification device for having anti-false function |
| CN110059607A (en) * | 2019-04-11 | 2019-07-26 | 深圳市华付信息技术有限公司 | Living body multiple detection method, device, computer equipment and storage medium |
| CN112395929A (en) * | 2019-08-19 | 2021-02-23 | 扬州盛世云信息科技有限公司 | Face living body detection method based on infrared image LBP histogram characteristics |
| CN110580454A (en) * | 2019-08-21 | 2019-12-17 | 北京的卢深视科技有限公司 | Living body detection method and device |
| CN112651268A (en) * | 2019-10-11 | 2021-04-13 | 北京眼神智能科技有限公司 | Method and device for eliminating black and white photos in biopsy, and electronic equipment |
| CN112651268B (en) * | 2019-10-11 | 2024-05-28 | 北京眼神智能科技有限公司 | Method and device for eliminating black-and-white photo in living body detection and electronic equipment |
| CN111879724A (en) * | 2020-08-05 | 2020-11-03 | 中国工程物理研究院流体物理研究所 | Human skin mask identification method and system based on near infrared spectrum imaging |
| CN111879724B (en) * | 2020-08-05 | 2021-05-04 | 中国工程物理研究院流体物理研究所 | Human skin mask identification method and system based on near infrared spectrum imaging |
| CN117953591A (en) * | 2024-03-27 | 2024-04-30 | 中国人民解放军空军军医大学 | Intelligent limb rehabilitation assisting method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107798281B (en) | 2021-10-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107798281A (en) | A kind of human face in-vivo detection method and device based on LBP features | |
| CN110516576B (en) | Near-infrared living body face recognition method based on deep neural network | |
| CN104166841B (en) | The quick detection recognition methods of pedestrian or vehicle is specified in a kind of video surveillance network | |
| CN107066942A (en) | A kind of living body faces recognition methods and system | |
| CN102708361B (en) | Human face collecting method at a distance | |
| CN108985170A (en) | Transmission line of electricity hanger recognition methods based on Three image difference and deep learning | |
| CN106250936A (en) | Multiple features multithreading safety check contraband automatic identifying method based on machine learning | |
| CN107798279B (en) | Face living body detection method and device | |
| CN106407917A (en) | Dynamic scale distribution-based retinal vessel extraction method and system | |
| CN109670430A (en) | A kind of face vivo identification method of the multiple Classifiers Combination based on deep learning | |
| CN106023151B (en) | Tongue object detection method under a kind of open environment | |
| CN109858439A (en) | A kind of biopsy method and device based on face | |
| CN106886216A (en) | Robot automatic tracking method and system based on RGBD Face datections | |
| CN109697430A (en) | The detection method that working region safety cap based on image recognition is worn | |
| CN109635846A (en) | A kind of multiclass medical image judgment method and system | |
| CN109087286A (en) | A kind of detection method and application based on Computer Image Processing and pattern-recognition | |
| CN103034838A (en) | Special vehicle instrument type identification and calibration method based on image characteristics | |
| CN108416774A (en) | A kind of fabric types recognition methods based on fine granularity neural network | |
| CN102254188A (en) | Palmprint recognizing method and device | |
| CN105224921A (en) | A kind of facial image preferentially system and disposal route | |
| CN104123543A (en) | Eyeball movement identification method based on face identification | |
| CN105205437B (en) | Side face detection method and device based on contouring head verifying | |
| CN106845328A (en) | A kind of Intelligent human-face recognition methods and system based on dual camera | |
| CN106650623A (en) | Face detection-based method for verifying personnel and identity document for exit and entry | |
| CN106709438A (en) | Method for collecting statistics of number of people based on video conference |
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 | ||
| CP02 | Change in the address of a patent holder |
Address after: 071800 Beijing Tianjin talent home (Xincheng community), West District, Xiongxian Economic Development Zone, Baoding City, Hebei Province Patentee after: BEIJING EYECOOL TECHNOLOGY Co.,Ltd. Address before: 100085 20 / F, building 4, yard 1, shangdishi street, Haidian District, Beijing 2013 Patentee before: BEIJING EYECOOL TECHNOLOGY Co.,Ltd. |
|
| CP02 | Change in the address of a patent holder | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: A method and device for detecting human face in vivo based on LBP feature Effective date of registration: 20220614 Granted publication date: 20211008 Pledgee: China Construction Bank Corporation Xiongxian sub branch Pledgor: BEIJING EYECOOL TECHNOLOGY Co.,Ltd. Registration number: Y2022990000332 |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
| CB03 | Change of inventor or designer information |
Inventor after: Kong Yong Inventor after: Zhang Xiangde Inventor before: Zhang Xiangde |
|
| CB03 | Change of inventor or designer information |