CN108416291A - Face datection recognition methods, device and system - Google Patents
Face datection recognition methods, device and system Download PDFInfo
- Publication number
- CN108416291A CN108416291A CN201810184392.XA CN201810184392A CN108416291A CN 108416291 A CN108416291 A CN 108416291A CN 201810184392 A CN201810184392 A CN 201810184392A CN 108416291 A CN108416291 A CN 108416291A
- Authority
- CN
- China
- Prior art keywords
- face
- human face
- image data
- concave
- convex curvature
- 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
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000001514 detection method Methods 0.000 claims abstract description 50
- 238000007373 indentation Methods 0.000 claims abstract description 22
- 238000004422 calculation algorithm Methods 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 12
- 238000001914 filtration Methods 0.000 claims description 9
- 238000005260 corrosion Methods 0.000 claims description 5
- 230000007797 corrosion Effects 0.000 claims description 5
- 230000008569 process Effects 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 12
- 238000005516 engineering process Methods 0.000 description 6
- 238000012549 training Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000005286 illumination Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000001815 facial effect Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 108010076504 Protein Sorting Signals Proteins 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 238000005530 etching Methods 0.000 description 1
- 238000001727 in vivo Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000003340 mental effect Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 235000001968 nicotinic acid Nutrition 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/50—Depth or shape recovery
-
- 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
- G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of Face datection recognition methods, device and system, which includes：Human face region is detected from the two dimensional image of acquisition；The two-dimensional image data of the human face region and acquired depth image data are projected into the same coordinate system to obtain 3 d image data；Whole concave-convex curvature detection is carried out using the 3 d image data and multiple localized indentation convex curvatures detect, and when whole concave-convex curvature is in the first preset range, whether the local weighted concave-convex curvature that judgement is obtained by multiple localized indentation convex curvatures is in the second preset range, if the local weighted concave-convex curvature is in second preset range, judge the human face region for real human face.The present invention does not need client in entirely detection identification process and does additional interoperation, it is ensured that keeps higher detection discrimination simultaneously under the premise of high recall rate, avoids the detections such as picture face being identified as real face.
Description
Technical field
The present invention relates to artificial intelligence fields, in particular to a kind of Face datection recognition methods, device and system.
Background technology
Face datection identifies, is a kind of biological detection identification that the facial feature information based on people carries out identity detection identification
Technology.Image or video flowing containing face, and automatic detect and track face in the picture are acquired with video camera or camera,
And then the face to detecting carries out a series of the relevant technologies of face, usually also referred to as portrait detection identification, face detection are known
Not.
Because Face datection identification has the characteristics that direct, convenient, user is easy to be connect by user without any mental handicape
By to obtain extensive research and application.
Traditional Face datection identifies the camera and collecting device identified based on 2D Face datections, is normally based on a small amount of
Sample go prediction, it is assumed that and the program by writing judge the texture information of face, face size, two interocular distances etc.,
What is obtained is the information of two dimensional surface.However, existing Face datection identifying system still has some loopholes, pass through static photograph
The modes such as piece, face video cheat existing detection identifying system.
With the development of anti-spoofing means, In vivo detection technology is come into being.This technology is generally matched using instruction action
The mode of conjunction, such as face turn left, turn right, open one's mouth, blink.But it is brought by way of completing authentication user's cooperation
User experience effect is simultaneously bad.
To solve the above problems, a kind of novel camera is released in camera market --- binocular camera.Binocular camera
It is to utilize bionics principle, synchronous exposure image is obtained by calibrated dual camera, then calculates the two dimensional image of acquisition
The third dimension depth information of pixel.But for computer, the depth information obtained by two cameras is not
Especially accurate, because can be influenced by light and other factors, error is larger；In addition very to the performance requirement of computing unit
Height, this makes the difficulty of the commercialization of existing biocular systems, miniaturization larger.
Invention content
In view of the above problems, the present invention provides a kind of new Face datection recognition methods, Face datection identification device and
Face datection identifying system.
An embodiment of the invention provides a kind of Face datection recognition methods, including：
Human face region is detected from the two dimensional image of acquisition；
The two-dimensional image data of the human face region and acquired depth image data are projected into the same coordinate system
To obtain 3 d image data；
Whole concave-convex curvature detection is carried out using the 3 d image data and multiple localized indentation convex curvatures detect, and
When whole bumps curvature is in the first preset range, judgement is by the local weighted concave-convex curvature that multiple localized indentation convex curvatures obtain
It is no in the second preset range, if the local weighted concave-convex curvature is in second preset range, judge the people
Face region is real human face.
In above-mentioned Face datection recognition methods, the human face region is detected using Adaboost algorithm.
In above-mentioned Face datection recognition methods, further include：Face Detection is carried out in the human face region to exclude
The human face region of flase drop.
In above-mentioned Face datection recognition methods, the local weighted concave-convex curvature utilizes left eye, right eye, mouth and nose
The weighted calculating of concave-convex curvature after obtain.
In above-mentioned Face datection recognition methods, further include：The height of nose is obtained using the 3 d image data
Data utilize the local weighted concave-convex curvature and described when the whole concave-convex curvature is in first preset range
Altitude information judges whether the human face region is real human face.
In above-mentioned Face datection recognition methods, further include：
In the case where being determined as real human face, the 3 d image data is compared with the image data to prestore with
Determine whether for registered members；
If it is determined that be registered members, then it is automatic to execute enabling or delivery operation；
If it is determined that being non-registered member, then user's registration member is reminded.
In above-mentioned Face datection recognition methods, before detecting the human face region, the two dimensional image is filtered
Wave processing is to remove noise；
Acquired depth image data have passed through pretreatment, and the pretreatment includes field compensation, medium filtering, corrosion
And expansion.
Another embodiment of the present invention provides a kind of Face datection identification device, including：
Human face region detection module detects human face region from the two dimensional image of acquisition；
3 d image data acquisition module, by the two-dimensional image data of the human face region and acquired depth image number
According to projection in the same coordinate system to obtain 3 d image data；
Face judgment module carries out whole concave-convex curvature detection and multiple localized indentation evaginations using the 3 d image data
Degree detection, and when whole concave-convex curvature is in the first preset range, judge the part obtained by multiple localized indentation convex curvatures
Whether the concave-convex curvature of weighting is in the second preset range, if the local weighted concave-convex curvature is in second preset range
When, then judge the human face region for real human face.
Another embodiment of the invention provides a kind of Face datection identifying system, including：
Two dimensional image camera, for obtaining two dimensional image；
Depth camera, for obtaining depth image；
Computer equipment, including memory and processor, the memory are stored with computer program, in the processor
When executing the computer program, implement above-mentioned Face datection recognition methods.
The a further embodiment of the present invention provides a kind of computer readable storage medium, is stored with above-mentioned Face datection
The computer program used in identifying system.
The three-dimensional face detection identifying system of the present invention is on the basis of two-dimensional detection identifies in detection and cognitive phase
Depth information is incorporated, to not need client in entirely detection identification process and do additional interoperation, it is ensured that in height
Higher detection discrimination is kept simultaneously under the premise of recall rate, avoids the detections such as picture face being identified as real face.
Description of the drawings
In order to illustrate more clearly of technical scheme of the present invention, letter will be made to attached drawing needed in the embodiment below
It singly introduces, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as to the present invention
The restriction of protection domain.
Fig. 1 shows the flow chart of one embodiment of the Face datection recognition methods of the present invention.
Fig. 2 shows the structural schematic diagrams of one embodiment of the Face datection identification device of the present invention.
Specific implementation mode
Below in conjunction with attached drawing in the embodiment of the present invention, technical solution in the embodiment of the present invention carries out clear, complete
Ground describes, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Usually exist
The component of the embodiment of the present invention described and illustrated in attached drawing can be arranged and be designed with a variety of different configurations herein.Cause
This, the detailed description of the embodiment of the present invention to providing in the accompanying drawings is not intended to limit claimed invention below
Range, but it is merely representative of the selected embodiment of the present invention.Based on the embodiment of the present invention, those skilled in the art are not doing
The every other embodiment obtained under the premise of going out creative work, shall fall within the protection scope of the present invention.
Embodiment 1
Fig. 1 shows the Face datection recognition methods of the present embodiment, in the face detection recognition method, including following step
Suddenly：
In step s 110, human face region is detected from the two dimensional image of acquisition.Two dimensional image can utilize common coloured silk
Color camera shoots the color image of personage, then, human face region detection is carried out to color image.
The method for detecting human face, template matching method and adaboost that feature based may be used in the detection of human face region are calculated
Method.The method for detecting human face of feature based includes overall profile method, Face Detection method and organ distribution；Template matching method includes
Mosaic map mosaic method (also known as mosaic method), predetermined template matching method and deforming template method.
Face datection algorithm based on the colour of skin meets rate height, in non-face area of skin color (trick, neck etc.) and background
Class area of skin color easy to produce erroneous judgement, while single fixed threshold value makes this algorithm robustness poor, be not suitable with illumination,
The environment of the extraneous factors such as shade variation, therefore the limitation of this algorithm, in general, can be only applied to the colour of skin under simple background
Detection.
The advantages of Face datection algorithm based on Gabor+BP neural networks is that the algorithm understands simple, learning ability of getting up
By force, calculation amount is small, has stronger noise resisting ability, can improve accuracy rate by expanding training sample set, but there is instruction
Practicing time length causes convergence rate slow, and learning efficiency is low, the choosing without theoretical direction to middle layer number and each middle layer node number
The problem of taking, learning to remember old sample well while new samples.
It is found through experiments that, the Face datection algorithm based on adaboost has certain rotation angle to the crowd of the different colours of skin
The face and illumination variation of degree have higher robustness, and verification and measurement ratio is high, and false drop rate is low.Therefore, it preferably adopts in this application
The detection of human face region is carried out with adaboost algorithms.
Adaboost is a kind of alternative manner, and core concept is to train the same weak typing for different training sets
Device constitutes a final strong classifier then the weak classifier set obtained on different training sets.
One Weak Classifier h (x, f, p, θ) is by child window image x, a feature f, the p and threshold value for indicating sign of inequality direction
θ is formed.The effect of P is the direction of majorization inequality so that inequality is all<Number, form is convenient.To each feature f, training one
A Weak Classifier h (x, f, p, θ) is just to determine the optimal threshold of f so that this Weak Classifier h (x, f, p, θ) is to all training
The error in classification of sample is minimum.To each feature, its feature to all a kind of samples (face is non-face) is calculated
The average value of value finally obtains average Distribution value of all features to all a kind of samples.
The pattern feature very abundant for including in facial image, such as histogram feature, color characteristic, template characteristic, structure
Feature and Haar features etc..Human face region detection is exactly that information useful among these is picked out, and is realized using these features
Human face region detects.
Above-mentioned Adaboost learning algorithms are used based on features above, are chosen using Adaboost algorithm during Face datection
The rectangular characteristic (Weak Classifier) of face can be represented by selecting some most, and Weak Classifier is configured to one in the way of Nearest Neighbor with Weighted Voting
A strong classifier, then the obtained several strong classifiers of training are composed in series the cascade filtering of a cascade structure, effectively
Improve the detection speed of grader.
Applicant have discovered that adaboost algorithms are more than 30 degree to image pitching waves and is tilted beyond 45 degree polygonal
It is not fine to spend Face datection effect, and the Face datection based on the colour of skin calculates algorithm and realizes that simply, calculating speed is fast, verification and measurement ratio
Height, it is insensitive to the face and expression shape change of multi-angle oblique.Therefore, in the present invention, it is preferred to utilizing adaboost algorithms
After carrying out human face region detection, further human face region is detected using Face Detection method, to exclude the face of flase drop
Region.Thus, it is possible to further enhance the accuracy of human face region detection.
May include establishing complexion model, similarity analysis and binaryzation based on Face Detection human face region.Establish the colour of skin
The process of model includes that RGB color is mapped to YCbCr color spaces, and YCbCr color spaces are a kind of colour model, wherein
Y refers to luminance information, and Cb and Cr indicate chrominance information.RGB models and the transformational relation in the spaces YcbCr are as follows：
R=Y+1.402 (Cr-0.5)
G=Y-0.3441 (Cb-0.5) -0.7141 (Cr-0.5)
B=Y+1.772 (Cb-0.5)
YCbCr color spaces can also separate coloration and luminance information, compare HSV space, and computation complexity is told somebody what one's real intentions are,
And in YCbCr space, the distribution of the colour of skin is more closely.
Similarity calculation refer to YCbCr color spaces normalize chroma histogram after, it is assumed that the colour of skin meets dimensional Gaussian
Model M=(m, C)；Wherein m is the mean value of coloration, i.e. m=E (x), C are the covariance matrix of coloration, pass through this Gauss model
Detect any one pixel whether be the colour of skin probability calculation formula it is as follows：
Here S is exactly covariance matrix, and μ is mean value or variance.Binaryzation be after obtaining face complexion probability graph, into
The processing of row binarization segmentation, to obtain face complexion bianry image, excludes the human face region of flase drop.
Before detecting the human face region, preferably the two dimensional image is filtered to remove noise.Filtering is gone
Make an uproar and exactly the noise in image removed, these noises may be the imaging of object caused by illumination or environment, device etc. and
It is in kind variant.Filtering out noise the methods of Gaussian filter can be used to carry out.
Further, it is also possible to carry out luminance compensation to two dimensional image, due to the brightness of illumination effect image, luminance compensation here
Gary World algorithms can be used, rapid color compensation, i.e., following formula are carried out to RGB image：S=av1/av2；F=f1*S；
Wherein av1 is the average value of criterion brightness image R, G, B, and av2 is the average value of original input image R, G, B, and f1 is that former input is schemed
The pixel value of picture, f are the pixel value of output image after luminance compensation.
In the step s 120, by the two-dimensional image data of the human face region and acquired depth image data project to
To obtain 3 d image data in the same coordinate system.
Depth image can utilize depth camera shooting to obtain.Depth camera may include infrared sensor and infrared hair
Emitter.The human face region detected in step S110 is projected in the coordinate of depth data to obtain three-dimension space image.Two
Dimension image and depth image are obtained by two different sensors, the coordinate system being respectively imaged be it is different, can will be it
All project under the same world coordinate system (physical coordinates system), the plane of such coloured image and the depth of depth image
Information, which combines, obtains 3 d image data, including three dimensional space coordinate and numerical value.
It is projected to before world coordinate system by depth image data, the image for preferably obtaining depth camera carries out pre-
Processing, such as at least one of field compensation, medium filtering, corrosion and expansion processing can be carried out.
Depth image may cause partial data to lose due to light or electric current etc., it is therefore desirable to compensate.Due to
There is distortion in camera imaging, the image quality height near optical axis is poor in edge quality, so compensation is divided into four regions, from
Central point nearby compensates around, is compensated with field mean value or field value.
Medium filtering to impulsive noise have it is good filter out effect, especially while filtering out noise, letter can be protected
Number edge, be allowed to not be blurred.Median filter method：To a digital signal sequences xj (- ∞<j<∞) it is filtered
When, it first has to define the L long windows that a length is odd number, L=2N+1, N are positive integer.Some moment is located at, in window
Sample of signal be x (i-N) ..., x (i) ..., x (i+N), wherein x (i) be positioned at window center sample of signal value.To this L
For a sample of signal value by after being ranked sequentially from small to large, intermediate value, the sample value at i is just defined as the output of medium filtering
Value.
Corrosion is that the region of gray value small (being exactly visually than dark) is enhanced extension, and it is brighter to be mainly used to removal
Noise.Expansion is that the region of gray value big (being exactly visually brighter) is enhanced extension, be mainly used to connection Similar color or
The region of intensity.The key concept that corrosion is related to expansion is exactly core, can also be referred to as template or mask.Core has
Several important attributes, shape (round, rectangular, cross is even oval), size (3x3,5x5 etc.) and reference point.It is most
In the case of, template it is fairly simple be a circle, size 3x3, reference point is in the center of circle.After determining core, just and original image
Carry out convolution.The gray value of each pixel of original image is equal to using the pixel as reference point, in the range of core covers, all pictures
The minimum gray value of member, this completes etching operations.Expansive working is exactly to be maximized.
In step s 130, concave-convex curvature detection is carried out to the 3 d image data.To in the 3 d image data
Depth data A asks convolution to be then maximized characterization spatial point respectively in horizontal vertical direction.
Horizontal convolution operatorVertical convolution operatorAfter processing
Depth image data GMax=Max (Cx*A, Cy*A).
Face distance D is estimated by depth data Aface, according to DfaceGenerate threshold value Tthreshold, using TthresholdTwo-value
Change GMaxObtain Bface, B is counted in conjunction with depth data AfaceWhole concave-convex curvature Sface(non-zero ratio).In whole concave-convex curvature
SfaceIn the first preset range, it is determined that the human face region is true face.
Then, it is determined that whether the local weighted concave-convex curvature obtained by multiple localized indentation convex curvatures is in the second preset range
It is interior, if the local weighted concave-convex curvature is in second preset range, judge the human face region for real human face.
The detection of each localized indentation convex curvature can also use above-mentioned similar method.Such as carrying out nose bumps curvature
Detection when, can be from BfaceThe data B of middle extraction nosenose, counted about the depth data of nose in conjunction in depth data A
BnoseNose bumps curvature Snose.Similar method may be used in the detection of the localized indentation convex curvature of face, eyes, cheek etc..
For example, the detection of localized indentation convex curvature can be directed to left eye, right eye, the left side of face, the right side of face, the concave-convex curvature of nose progress
Detection.
Acquisition is weighted by multiple localized indentation convex curvatures in local weighted bumps curvature.Formula can be usedObtain local weighted concave-convex curvature；WiFor the weight of each localized indentation convex curve；SiFor each localized indentation evagination
Degree, such as S0To S4Can be the concave-convex curvature of left eye, right eye, the left side of face, the right side of face, nose respectively.Certainly, i
Maximum value may be other natural numbers n, such as can not have to face being divided into the left and right sides, at this point, i is 3.If local weighted
When concave-convex curvature is in the second preset range, then judge the human face region for real human face.
In the present invention, the altitude information of nose is further preferably obtained using the 3 d image data, it is recessed in the entirety
When convex curvature is in first preset range, the people is judged using the local weighted concave-convex curvature and the altitude information
Whether face region is real human face.For example, when whole concave-convex curvature is in first preset range, using following formula come
Judge whether the human face region is real human face：
WiFor the weight of each localized indentation convex curve；SiFor each localized indentation convex curvature；WnoseFor the weight of nose height；
YnoseIt is 1 or 0, it can be according to calculating nose height value HnoseWhether determine whether within a predetermined range for nose, if nose
It is then 1, if not nose is then 0.Above-mentioned each weight can be obtained by machine learning or neural network learning.The height of nose
Degrees of data can utilize 3 d image data to obtain, and Color images detecting is to the region that may be face, then from human face region
Detect nose, the color coordinate of face is projected to three dimensions, from three dimensional space coordinate obtain altitude information, ask minimum point and
Peak is exactly nose height.
Judged whether in preset range according to Face values to determine whether be true face, and the preset range of Face with
Actual environment is related, for example, in face distance DfaceWhen (with a distance from camera) is 1m, Face is more than 0.55, then shows institute
It is true face to state human face region.Those skilled in the art can determine the Face values under different distances according to experiment
Preset range.
It, can in the case where being determined as face when above-mentioned Face datection recognition methods is for such as unmanned sale shop
Compareing to determine if the 3 d image data and the image data to prestore for registered members；If it is determined that for registration
Member then automatically opens shop door when user enters unmanned shop, or is paid when user settles accounts；If it is determined that not being note
Volume member, then can prompt user to register.
Face datection recognition methods according to the present invention can extract face area roughly in detection-phase by Adaboost
Domain, but also the face of flase drop can be further excluded by Face Detection.Face after Face Detection is put into three dimensional detection,
Three dimensional detection can be detected according to the entirety and localized indentation convex curvature of face to be judged.
Three dimensional detection is identified by face's Global Information of three-dimensional nerve network extraction people and crucial point feature, with three-dimensional
Information characteristics characterize true face, so as to effectively exclude dummies' face such as plane picture, improve the accuracy rate of detection identification.
Therefore, Face datection recognition methods of the invention can effectively detect true face.
Embodiment 2
Fig. 2 shows the Face datection identification devices 200 of the present embodiment.In the Face datection identification device 200, packet
It includes：Human face region detection module 210 detects human face region from the two dimensional image of acquisition；3 d image data acquisition module
220, the two-dimensional image data of the human face region and acquired depth image data are projected into the same coordinate system to obtain
Obtain 3 d image data；Face judgment module 230 is detected and multiple using the whole concave-convex curvature of 3 d image data progress
Localized indentation convex curvature detects, and when whole concave-convex curvature is in the first preset range, judges by multiple localized indentation convex curvatures
Whether the local weighted concave-convex curvature obtained is in the second preset range, if the local weighted concave-convex curvature is pre- described second
When determining in range, then judge the human face region for real human face.
Above-mentioned Face datection identification device corresponds to the Face datection recognition methods of embodiment 1.It is any in embodiment 1
Option is also applied for the present embodiment, and I will not elaborate.
The present invention also provides a kind of Face datection identifying systems, including：Two dimensional image camera, for obtaining X-Y scheme
Picture；Depth camera, for obtaining depth image；Computer equipment, including memory and processor, the memory are stored with
Computer program implements above-mentioned Face datection recognition methods when the processor executes the computer program.X-Y scheme
As camera and depth camera can integrate composition three-dimensional camera.For example, the three-dimensional camera may include colour
Sensor, infrared sensor, Infrared laser emission device and picture processing chip.
Computer equipment, including memory and processor, the memory are stored with computer program, in the processor
When executing the computer program, to make computer equipment execute above-mentioned Face datection recognition methods or above-mentioned Face datection
The function of modules in identification device.
Memory may include storing program area and storage data field, wherein storing program area can storage program area, at least
Application program (such as sound-playing function, image player function etc.) needed for one function etc.；Storage data field can store root
Created data (such as audio data, phone directory etc.) etc. are used according to mobile terminal.In addition, memory may include high speed
Random access memory, can also include nonvolatile memory, a for example, at least disk memory, flush memory device or
Other volatile solid-state parts.
The present embodiment additionally provides a kind of computer storage media, is used for storing in above-mentioned Face datection identifying system
Computer program.
In several embodiments provided herein, it should be understood that disclosed device and method can also pass through
Other modes are realized.The apparatus embodiments described above are merely exemplary, for example, the flow chart in attached drawing and structure
Figure show the device of multiple embodiments according to the present invention, method and computer program product system frame in the cards
Structure, function and operation.In this regard, each box in flowchart or block diagram can represent a module, section or code
A part, the part of the module, section or code includes one or more for implementing the specified logical function
Executable instruction.It should also be noted that in the realization method as replacement, the function of being marked in box can also be to be different from
The sequence marked in attached drawing occurs.For example, two continuous boxes can essentially be basically executed in parallel, they are sometimes
It can execute in the opposite order, this is depended on the functions involved.It is also noted that in structure chart and/or flow chart
The combination of each box and the box in structure chart and/or flow chart can use the special of function or action as defined in executing
Hardware based system realize, or can realize using a combination of dedicated hardware and computer instructions.
In addition, each function module or unit in each embodiment of the present invention can integrate and to form an independence
Part, can also be modules individualism, can also two or more modules be integrated to form an independent part.
It, can be with if the function is realized and when sold or used as an independent product in the form of software function module
It is stored in a computer read/write memory medium.Based on this understanding, technical scheme of the present invention is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be expressed in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be intelligence
Can mobile phone, personal computer, server or network equipment etc.) execute each embodiment the method for the present invention whole or
Part steps.And storage medium above-mentioned includes：USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory),
Random access memory (RAM, Random Access Memory), magnetic disc or CD etc. are various can to store program code
Medium.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.
Claims (10)
1. a kind of Face datection recognition methods, which is characterized in that including：
Human face region is detected from the two dimensional image of acquisition；
The two-dimensional image data of the human face region and acquired depth image data are projected into the same coordinate system to obtain
Obtain 3 d image data；
Whole concave-convex curvature detection is carried out using the 3 d image data and multiple localized indentation convex curvatures detect, and in entirety
When concave-convex curvature is in the first preset range, judgement by multiple localized indentation convex curvatures obtains it is local weighted bumps curvature whether
In second preset range, if the local weighted concave-convex curvature is in second preset range, the face area is judged
Domain is real human face.
2. Face datection recognition methods according to claim 1, which is characterized in that detect institute using adaboost algorithms
State human face region.
3. Face datection recognition methods according to claim 2, which is characterized in that further include：In the human face region
Face Detection is carried out to exclude the human face region of flase drop.
4. Face datection recognition methods according to claim 1, which is characterized in that the local weighted concave-convex curvature utilizes
Left eye, right eye, mouth and nose the weighted calculating of concave-convex curvature after obtain.
5. Face datection recognition methods according to claim 1, which is characterized in that further include：Utilize the 3-D view
The altitude information of data acquisition nose utilizes the part when the whole concave-convex curvature is in first preset range
The concave-convex curvature of weighting and the altitude information judge whether the human face region is real human face.
6. Face datection recognition methods according to claim 1, which is characterized in that further include：
In the case where being determined as real human face, the 3 d image data is compared to determine with the image data to prestore
Whether it is registered members；
If it is determined that be registered members, then it is automatic to execute enabling or delivery operation；
If it is determined that being non-registered member, then user's registration member is reminded.
7. Face datection recognition methods according to claim 1, it is characterised in that：
Before detecting the human face region, the two dimensional image is filtered to remove noise；
Acquired depth image data have passed through pretreatment, and the pretreatment includes field compensation, medium filtering, corrosion and swollen
It is at least one of swollen.
8. a kind of Face datection identification device, which is characterized in that including：
Human face region detection module detects human face region from the two dimensional image of acquisition；
3 d image data acquisition module throws the two-dimensional image data of the human face region and acquired depth image data
To obtain 3 d image data in shadow to the same coordinate system；
Face judgment module carries out whole concave-convex curvature detection using the 3 d image data and multiple localized indentation convex curvatures is examined
It surveys, and when whole concave-convex curvature is in the first preset range, judgement is obtained local weighted by multiple localized indentation convex curvatures
Whether concave-convex curvature is in the second preset range, if the local weighted concave-convex curvature is in second preset range,
Judge the human face region for real human face.
9. a kind of Face datection identifying system, which is characterized in that including：
Two dimensional image camera, for obtaining two dimensional image；
Depth camera, for obtaining depth image；
Computer equipment, including memory and processor, the memory are stored with computer program, are executed in the processor
When the computer program, implement the Face datection recognition methods described in any one of claim 1-7.
10. a kind of computer readable storage medium, which is characterized in that it is stored with the identification of the Face datection described in claim 9
The computer program used in system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810184392.XA CN108416291B (en) | 2018-03-06 | 2018-03-06 | Face detection and recognition method, device and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810184392.XA CN108416291B (en) | 2018-03-06 | 2018-03-06 | Face detection and recognition method, device and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108416291A true CN108416291A (en) | 2018-08-17 |
| CN108416291B CN108416291B (en) | 2021-02-19 |
Family
ID=63129978
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810184392.XA Expired - Fee Related CN108416291B (en) | 2018-03-06 | 2018-03-06 | Face detection and recognition method, device and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108416291B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109934195A (en) * | 2019-03-21 | 2019-06-25 | 东北大学 | A kind of anti-spoofing three-dimensional face identification method based on information fusion |
| CN110309782A (en) * | 2019-07-02 | 2019-10-08 | 四川大学 | It is a kind of based on infrared with visible light biocular systems living body faces detection methods |
| CN111144183A (en) * | 2018-11-06 | 2020-05-12 | 天地融科技股份有限公司 | Risk detection method, device and system based on face concavity and convexity |
| CN111382592A (en) * | 2018-12-27 | 2020-07-07 | 杭州海康威视数字技术股份有限公司 | Living body detection method and apparatus |
| CN111898553A (en) * | 2020-07-31 | 2020-11-06 | 成都新潮传媒集团有限公司 | Method and device for distinguishing virtual image personnel and computer equipment |
| CN112149580A (en) * | 2020-09-25 | 2020-12-29 | 江苏邦融微电子有限公司 | Image processing method for distinguishing real human face from photo |
| CN112784661A (en) * | 2019-11-01 | 2021-05-11 | 宏碁股份有限公司 | Real face recognition method and real face recognition device |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1794265A (en) * | 2005-12-31 | 2006-06-28 | 北京中星微电子有限公司 | Method and device for distinguishing face expression based on video frequency |
| CN1975759A (en) * | 2006-12-15 | 2007-06-06 | 中山大学 | Human face identifying method based on structural principal element analysis |
| CN101630363A (en) * | 2009-07-13 | 2010-01-20 | 中国船舶重工集团公司第七○九研究所 | Rapid detection method of face in color image under complex background |
| US7670727B2 (en) * | 2003-12-09 | 2010-03-02 | Anvik Corporation | Illumination compensator for curved surface lithography |
| CN103246875A (en) * | 2013-05-09 | 2013-08-14 | 东南大学 | Three-dimensional facial recognition method based on elasticity matching of facial curves |
| CN103440479A (en) * | 2013-08-29 | 2013-12-11 | 湖北微模式科技发展有限公司 | Method and system for detecting living body human face |
| CN104951767A (en) * | 2015-06-23 | 2015-09-30 | 安阳师范学院 | Three-dimensional face recognition technology based on correlation degree |
| CN105022994A (en) * | 2015-06-30 | 2015-11-04 | 国网山东省电力公司日照供电公司 | Identity authentication method of network safety access of power system |
| CN105740778A (en) * | 2016-01-25 | 2016-07-06 | 北京天诚盛业科技有限公司 | Improved three-dimensional human face in-vivo detection method and device thereof |
| CN105844276A (en) * | 2015-01-15 | 2016-08-10 | 北京三星通信技术研究有限公司 | Face posture correction method and face posture correction device |
| CN105913263A (en) * | 2016-04-01 | 2016-08-31 | 曹龙巧 | Application method of face payment platform based on image detection |
| CN106530383A (en) * | 2016-11-01 | 2017-03-22 | 河海大学 | Human face rendering method based on Hermite interpolation neural network regression model |
| US20180033205A1 (en) * | 2016-08-01 | 2018-02-01 | Lg Electronics Inc. | Mobile terminal and operating method thereof |
-
2018
- 2018-03-06 CN CN201810184392.XA patent/CN108416291B/en not_active Expired - Fee Related
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7670727B2 (en) * | 2003-12-09 | 2010-03-02 | Anvik Corporation | Illumination compensator for curved surface lithography |
| CN1794265A (en) * | 2005-12-31 | 2006-06-28 | 北京中星微电子有限公司 | Method and device for distinguishing face expression based on video frequency |
| CN1975759A (en) * | 2006-12-15 | 2007-06-06 | 中山大学 | Human face identifying method based on structural principal element analysis |
| CN101630363A (en) * | 2009-07-13 | 2010-01-20 | 中国船舶重工集团公司第七○九研究所 | Rapid detection method of face in color image under complex background |
| CN103246875A (en) * | 2013-05-09 | 2013-08-14 | 东南大学 | Three-dimensional facial recognition method based on elasticity matching of facial curves |
| CN103440479A (en) * | 2013-08-29 | 2013-12-11 | 湖北微模式科技发展有限公司 | Method and system for detecting living body human face |
| CN105844276A (en) * | 2015-01-15 | 2016-08-10 | 北京三星通信技术研究有限公司 | Face posture correction method and face posture correction device |
| CN104951767A (en) * | 2015-06-23 | 2015-09-30 | 安阳师范学院 | Three-dimensional face recognition technology based on correlation degree |
| CN105022994A (en) * | 2015-06-30 | 2015-11-04 | 国网山东省电力公司日照供电公司 | Identity authentication method of network safety access of power system |
| CN105740778A (en) * | 2016-01-25 | 2016-07-06 | 北京天诚盛业科技有限公司 | Improved three-dimensional human face in-vivo detection method and device thereof |
| CN105913263A (en) * | 2016-04-01 | 2016-08-31 | 曹龙巧 | Application method of face payment platform based on image detection |
| US20180033205A1 (en) * | 2016-08-01 | 2018-02-01 | Lg Electronics Inc. | Mobile terminal and operating method thereof |
| CN106530383A (en) * | 2016-11-01 | 2017-03-22 | 河海大学 | Human face rendering method based on Hermite interpolation neural network regression model |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111144183A (en) * | 2018-11-06 | 2020-05-12 | 天地融科技股份有限公司 | Risk detection method, device and system based on face concavity and convexity |
| CN111144183B (en) * | 2018-11-06 | 2024-05-28 | 天地融科技股份有限公司 | Risk detection method, device and system based on face concave-convex degree |
| CN111382592B (en) * | 2018-12-27 | 2023-09-29 | 杭州海康威视数字技术股份有限公司 | Living body detection method and apparatus |
| CN111382592A (en) * | 2018-12-27 | 2020-07-07 | 杭州海康威视数字技术股份有限公司 | Living body detection method and apparatus |
| CN109934195A (en) * | 2019-03-21 | 2019-06-25 | 东北大学 | A kind of anti-spoofing three-dimensional face identification method based on information fusion |
| CN110309782B (en) * | 2019-07-02 | 2022-05-03 | 四川大学 | Living body face detection method based on infrared and visible light binocular system |
| CN110309782A (en) * | 2019-07-02 | 2019-10-08 | 四川大学 | It is a kind of based on infrared with visible light biocular systems living body faces detection methods |
| CN112784661A (en) * | 2019-11-01 | 2021-05-11 | 宏碁股份有限公司 | Real face recognition method and real face recognition device |
| CN112784661B (en) * | 2019-11-01 | 2024-01-19 | 宏碁股份有限公司 | Real face recognition method and real face recognition device |
| CN111898553A (en) * | 2020-07-31 | 2020-11-06 | 成都新潮传媒集团有限公司 | Method and device for distinguishing virtual image personnel and computer equipment |
| CN111898553B (en) * | 2020-07-31 | 2022-08-09 | 成都新潮传媒集团有限公司 | Method and device for distinguishing virtual image personnel and computer equipment |
| CN112149580A (en) * | 2020-09-25 | 2020-12-29 | 江苏邦融微电子有限公司 | Image processing method for distinguishing real human face from photo |
| CN112149580B (en) * | 2020-09-25 | 2024-05-14 | 江苏邦融微电子有限公司 | Image processing method for distinguishing real face from photo |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108416291B (en) | 2021-02-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108038456B (en) | Anti-deception method in face recognition system | |
| CN108416291A (en) | Face datection recognition methods, device and system | |
| KR102561723B1 (en) | System and method for performing fingerprint-based user authentication using images captured using a mobile device | |
| CN105205480B (en) | Human-eye positioning method and system in a kind of complex scene | |
| JP5629803B2 (en) | Image processing apparatus, imaging apparatus, and image processing method | |
| US8345936B2 (en) | Multispectral iris fusion for enhancement and interoperability | |
| CN101142584B (en) | Method for facial features detection | |
| JP4743823B2 (en) | Image processing apparatus, imaging apparatus, and image processing method | |
| Kashem et al. | Face recognition system based on principal component analysis (PCA) with back propagation neural networks (BPNN) | |
| WO2019137178A1 (en) | Face liveness detection | |
| CN106446872A (en) | Detection and recognition method of human face in video under low-light conditions | |
| WO2016010724A1 (en) | Multispectral eye analysis for identity authentication | |
| CN109858439A (en) | A kind of biopsy method and device based on face | |
| WO2016010720A1 (en) | Multispectral eye analysis for identity authentication | |
| WO2016010721A1 (en) | Multispectral eye analysis for identity authentication | |
| CN105426843B (en) | Single-lens lower palm vein and palm print image acquisition device and image enhancement and segmentation method | |
| JP2000003452A (en) | Method for detecting face surface in digital picture, its detecting device, picture judging method, picture judging device and computer readable record medium | |
| WO2009078957A1 (en) | Systems and methods for rule-based segmentation for objects with full or partial frontal view in color images | |
| CN108629262A (en) | Iris identification method and related device | |
| CN106650606A (en) | Matching and processing method of face image and face image model construction system | |
| CN109034133A (en) | A kind of face identification method and device | |
| Hebbale et al. | Real time COVID-19 facemask detection using deep learning | |
| EP3885982A2 (en) | Hand biometrics system and method using digital fingerprints | |
| TWI427545B (en) | Face recognition method based on sift features and head pose estimation | |
| Parikh et al. | Effective approach for iris localization in nonideal imaging conditions |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210219 |