CN106682578A - Human face recognition method based on blink detection - Google Patents
Human face recognition method based on blink detection Download PDFInfo
- Publication number
- CN106682578A CN106682578A CN201611039329.4A CN201611039329A CN106682578A CN 106682578 A CN106682578 A CN 106682578A CN 201611039329 A CN201611039329 A CN 201611039329A CN 106682578 A CN106682578 A CN 106682578A
- Authority
- CN
- China
- Prior art keywords
- image
- user
- facial image
- blink
- face
- 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 Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention provides a human face recognition method based on blink detection. The method comprises the steps: taking a picture of a user through a camera, and detecting a human face image from the picture of the user through a human face detector; carrying out the matching of the human face image with a pre-stored registered human face image through a human face matching algorithm, and determining that the human face is registered when the similarity of the matching result is greater than a set matching threshold value; detecting whether a plurality of user images have the blink phenomenon or not through a blink detection algorithm, and determining that the user images are the images of a living body if the user images have the blink phenomenon. The method can detect whether the detected object is a picture, a sculpture or a living body well through blink detection. Compared with an ASM+Canny blind detection method which is higher in requirements for light, the method carries out the differential processing of the image through the morphology, and can be well suitable for a weak light detection condition.
Description
Technical field
The present invention relates to technical field of face recognition, more particularly to a kind of face identification method based on blink detection.
Background technology
Face recognition technology is the computer application investigative technique of a hot topic, and it belongs to biometrics identification technology, is led to
It is individual to distinguish organism to cross the biological characteristic to organism itself.Recognition of face is the most natural, visual a kind of raw
Thing identification mode, meets the physiological vision custom of mankind itself.With the development of face recognition technology, accuracy, can
All will improve a lot by property.There is now application widely, including public security, safety, customs, finance, army, airport,
The multiple important industries of security protection etc. and field, and the commercial market such as intelligent entrance guard, work attendance, market application foreground is boundless.
Face recognition technology mainly includes several parts：Face datection, face matching, face tracking, In vivo detection etc..
There is more ripe algorithm currently for each part, but for the In vivo detection under complex environment is then needed more
Comprehensive study form further modified algorithm together.
Face recognition scheme of the prior art mainly includes following several：
1. recognition of face：
A) complexion model method：It is distributed the rule of Relatively centralized in color space to be detected according to the looks colour of skin；
B) sample learning method：Using the method for artificial neural network in pattern-recognition, i.e., by opposite as sample sets and non-
The study of image surface sample sets produces grader；
C) face identification method of feature based face (PCA)：All image surface set are considered as into an image surface subspace, and base
Image surface is judged whether in detection sample and its distance between the projection of subspace.
2. blink detection：
ASM+Canny：By the use of ASM (Active Shape Model active shape models) algorithms as the inspection of human eye area
Survey method, recycles Canny operators to calculate the edge contour of human eye, is blinked by the Distance Judgment up and down of human eye edge contour.
3. face matching：
Feature vector method：First determine the category such as size, position, the distances of image surface face profile such as eye iris, the wing of nose, the corners of the mouth
Property, their geometric feature is then calculated again, and these characteristic quantities form a characteristic vector for describing the image surface.
The shortcoming of above-mentioned face recognition scheme of the prior art is：
1. most of face recognition software of the prior art also can be by checking using the photo of registrant, can not area
It is living person or photo to divide, and this greatly reduces the security of recognition of face..
2. the blink detection method of ASM+Canny, the requirement to light is very high.Under the low light level, locating effects of the ASM to human eye
It is to human eye rim detection effect and bad with Canny.
3. for can not very well recognize face under low light condition.
4. existing algorithm is not high in different photoenvironment human face matching degrees.
The content of the invention
The embodiment provides a kind of face identification method based on blink detection, to realize well to user
Image carry out recognition of face.
To achieve these goals, this invention takes following technical scheme.
A kind of face identification method based on blink detection, including：
The image of user is shot by camera, face is detected from the image of the user by human-face detector
Image；
The facial image is matched with the chartered facial image for prestoring by face matching algorithm,
When the result of the matching is that similarity is more than the matching threshold of setting, it is determined that the facial image has been registered；
Detected by blink detection algorithm in the image of multiple users with the presence or absence of blink, when detecting that presence blinks
Eye, it is determined that the image of the user is from live body.
Further, the described image that user is shot by camera, by human-face detector from the figure of the user
Facial image is detected as in, including：
Using Adaboost method construct human-face detectors, the human-face detector for screening type cascade classifier, level
Each node of connection arranges on each node judgment threshold, position node bag more rearward according to cascade sort algorithm arrangement
Containing more Weak Classifiers；
The image of the user of each frame that camera is shot is input into the human-face detector, the human-face detector
The image of the user of each frame is detected successively, from the image of the user facial image is detected.
Further, it is described by face matching algorithm by the facial image and the chartered people for prestoring
Face image is matched, when the result of the matching is that similarity is more than the matching threshold of setting, it is determined that the face figure
As having registered, including：
After facial image is detected in the image from the user of present frame, list will be converted into after the facial image cutting
Channel image, by the registered facial image for prestoring single channel image is converted into, and described two single channel images are made into 4
Plant histogram to compare, draw 4 Similarity values：CV_COMP_CHISQR card sides, CV_COMP_BHATTACHARYYA distances, CV_
The COMP_CORREL degrees of correlation, CV_COMP_INTERSECT intersects coefficient；
Final similarity between described two single channel images is determined according to 4 Similarity values, when described final
Similarity is more than similarity threshold set in advance, then judge that the match is successful for described two single channel images, determines the face
Image has been registered.
Further, described method also includes：
When the final similarity is not more than similarity threshold set in advance, then described two single channel images are judged
With failure, the brightness and/or contrast to the registered facial image is adjusted, and regenerates the face figure after adjustment
The single channel image of picture；
Two single channel images are matched again according to above-mentioned matching process, when two single channel images, the match is successful, really
The fixed facial image has been registered；When two single channel images, it fails to match, then continue to the registered facial image
Brightness and/or contrast be adjusted, continuation two single channel images are matched again according to above-mentioned matching process, work as matching
The number of times of failure is more than setting numerical value, it is determined that the facial image is not registered.
Further, described method also includes：
Brightness and/or contrast to the registered facial image is adjusted using an Operators Algorithm, and formula is such as
Under：
G (i, j)=α f (i, j)+β
Parameter alpha>0, it is gain, for adjusting contrast, β is offset parameter, and for adjusting brightness, f (i, j) is source images
Pixel, g (i, j) is output image pixel, and i, j represents that pixel is arranged positioned at the i-th row jth.
Further, with the presence or absence of blink in the described image that multiple users are detected by blink detection algorithm,
There is blink when detecting, it is determined that the image of the user from live body, including：
The image of the user of each frame is converted into bianry image, the searching component pair in the bianry image, root
According to the restrictive condition of the width of eye element set in advance, height, horizontal range and vertical range, to the component pair for searching out
Filtered, acquisition meets the eye element pair of filter condition, and obtains region of search of the eye element to place；
According to the component for meeting filter condition to building eye template, calculated using normalizated correlation coefficient algorithm
Similar value between the eye template and the region of search, the region of search in the continuous image of two framed users in front and back
With the difference between the similar value between the eye template more than the judgment threshold for setting, it is determined that in the image of the user
Comprising blink, the image of the user is from live body.
The technical scheme provided by embodiments of the invention described above can be seen that provided in an embodiment of the present invention based on blink
By detection blink, can well distinguish detection is photo, sculpture or live body to the face identification method of detection.Compared to ASM
The blink detection method of+Canny to the exigent shortcoming of light (under the low light level, locating effects and Canny of the ASM to human eye
It is to human eye rim detection effect and bad), and the present invention carries out difference processing to image using morphology, can well adapt to weak
Light testing conditions.
The additional aspect of the present invention and advantage will be set forth in part in the description, and these will become from the following description
Obtain substantially, or recognized by the practice of the present invention.
Description of the drawings
In order to be illustrated more clearly that the technical scheme of the embodiment of the present invention, below will be to use needed for embodiment description
Accompanying drawing be briefly described, it should be apparent that, drawings in the following description are only some embodiments of the present invention, for this
For the those of ordinary skill of field, without having to pay creative labor, can be obtaining other according to these accompanying drawings
Accompanying drawing.
Fig. 1 is a kind of handling process of face identification method based on blink detection provided in an embodiment of the present invention；
Fig. 2 is a kind of principle schematic of the method for Face datection based on adaboost provided in an embodiment of the present invention；
Fig. 3 is a kind of schematic diagram of live body blink detection provided in an embodiment of the present invention.
Specific embodiment
Embodiments of the present invention are described below in detail, the example of the embodiment is shown in the drawings, wherein ad initio
Same or similar element is represented to same or similar label eventually or the element with same or like function.Below by
The embodiment being described with reference to the drawings is exemplary, is only used for explaining the present invention, and is not construed as limiting the claims.
Those skilled in the art of the present technique are appreciated that unless expressly stated, singulative " " used herein, "
It is individual ", " described " and " being somebody's turn to do " may also comprise plural form.It is to be further understood that arranging used in the specification of the present invention
Diction " including " refers to there is the feature, integer, step, operation, element and/or component, but it is not excluded that existing or adding
One or more other features, integer, step, operation, element, component and/or their group.It should be understood that when we claim unit
Part is " connected " or during " coupled " to another element, and it can be directly connected or coupled to other elements, or can also exist
Intermediary element.Additionally, " connection " used herein or " coupling " can include wireless connection or couple.Wording used herein
"and/or" includes one or more associated any cells for listing item and all combination.
Those skilled in the art of the present technique are appreciated that unless otherwise defined all terms used herein are (including technology art
Language and scientific terminology) have with art of the present invention in those of ordinary skill general understanding identical meaning.Should also
It is understood by, those terms defined in such as general dictionary should be understood that the meaning having with the context of prior art
The consistent meaning of justice, and unless defined as here, will not be with idealizing or excessively formal implication is explaining.
For ease of the understanding to the embodiment of the present invention, do by taking several specific embodiments as an example further below in conjunction with accompanying drawing
Explanation, and each embodiment does not constitute the restriction to the embodiment of the present invention.
By detection blink, that can well distinguish detection is photo or living person to the present invention.The present invention is poor using morphology
Partial image, can well adapt to Dim light measurement condition.The present invention is highly effective positive dough figurine using Harr+Adaboost methods
Face detecting method, it has very big advantage on algorithm speed, and the verification and measurement ratio of front face is very high.Carried out using the method
The model file that training is obtained, has all well and good effect to the Face datection under low light condition.The present invention utilizes a point operator, repeatedly
Change the brightness and contrast of registered human face photo, then carry out face matching checking, so as to the rate that improves that the match is successful.
A kind of handling process of face identification method based on blink detection provided in an embodiment of the present invention as shown in figure 1,
Including following process step：
Step 1, the image of authenticated user is stored in the server.
The image of authenticated user is registered in the server, the title set according to user is by authenticated use
The image at family is preserved in the server with forms such as jpg, and the login name of user can be stored in the registration table of TXT forms.When
Multiple user's registrations, can represent n-th facial image in registration table with Xn.
Step 2, the image that user is shot by camera, are detected by human-face detector from the image of the user
Facial image.
Using Adaboost method construct human-face detectors, the human-face detector for screening type cascade classifier, level
Each node of connection arranges on each node judgment threshold, position node bag more rearward according to cascade sort algorithm arrangement
Containing more Weak Classifiers；
The image of the user of each frame that camera is shot is input into the human-face detector, the human-face detector
The image of the user of each frame is detected successively, from the image of the user facial image is detected.
A kind of principle schematic of the method for Face datection based on adaboost that the present invention is provided is as shown in Fig. 2 profit
The human-face detector with Adaboost method constructs, using waterfall (Cascade) algorithm classification device is organized as the cascade of screening type
Grader, each node of cascade is the strong classifier that AdaBoost training is obtained.In each node of cascade, threshold value b is set,
So that nearly all face sample can be transferred through, and most non-face samples can not pass through.Node is arranged from simple to complex
Row, position node more rearward is more complicated, i.e., comprising more Weak Classifiers.Such energy minimization refusal image but during region
Amount of calculation, notifies the high detection rate and low reject rate for ensureing grader.It is 99.9% for example in discrimination, reject rate is 50%
When, (99.9% face and 50% non-face can pass through), total discrimination of 20 nodes is：0.99920≈ 98%, and
False acceptance rate is only：0.520≈ 0.0001%.For the face that picture is detected, it is labeled with red frame.
Step 3, the facial image is entered with the chartered facial image for prestoring by face matching algorithm
Row matching, when the result of matching is that similarity is more than the matching threshold of setting, it is determined that the facial image has been registered.
After facial image is detected in the image from the user of present frame, being converted into single channel after the cutting of face part
Image, by the registered facial image for prestoring single channel image is converted into, and two single channel images are made into 4 kinds of histograms
Relatively, 4 Similarity values are drawn：CV_COMP_CHISQR card sides, CV_COMP_BHATTACHARYYA distances, CV_COMP_
The CORREL degrees of correlation, CV_COMP_INTERSECT intersects coefficient；
The final similarity between two single channel images is determined according to 4 Similarity values, when final similarity is more than pre-
The similarity threshold for first setting, then judge that the match is successful for two single channel images, determines that facial image has been registered.
Both are straight for CV_COMP_CHISQR (card side), CV_COMP_BHATTACHARYYA (Bhattacharyya distances)
The value that the comparison of square figure draws is less, and explanatory diagram picture is more similar, and CV_COMP_CORREL (degree of correlation), CV_COMP_
INTERSECT (intersecting coefficient) both histograms relatively show that value is bigger, and explanatory diagram picture is more similar.By this four similarities
Value can determine that final Similarity value.
A might as well be set, b, c, d represent respectively CV_COMP_CHISQR card sides, CV_COMP_BHATTACHARYYA distances, CV_
The COMP_CORREL degrees of correlation, the value that the intersecting four kinds of histograms of coefficient of CV_COMP_INTERSECT relatively draw
Work as a, at least one is less than or equal to 0.1 to b, and c, d at least one when being more than or equal to 0.9, be judged to two figures
90% is reached as similar；
By that analogy,
Work as a, at least one is less than or equal to 0.2 to b, and c, d at least one when being more than or equal to 0.8, be judged to two figures
80% is reached as similar.
When the final similarity is not more than similarity threshold set in advance, then described two single channel images are judged
With failure.Because the method that this histogram compares facial image is affected by the brightness of picture, to registered facial image
Brightness and/or contrast are adjusted, and regenerate the single channel image of the facial image after adjustment.According to above-mentioned matching process
Two single channel images are matched again, when two single channel images, the match is successful, determines that the facial image has been registered；When
It fails to match for two single channel images, then continue the brightness to the registered facial image and/or contrast is adjusted,
Continuation again matches two single channel images according to above-mentioned matching process, when the number of times that it fails to match is more than numerical value is set, then
Determine that the facial image is not registered.
Brightness and/or contrast to registered facial image is adjusted using an Operators Algorithm, and formula is as follows：
G (i, j)=α f (i, j)+β
Two parameter alphas>0 and β is commonly referred to as gain and offset parameter.The two parameters are controlling respectively contrast and bright
Degree.F (i, j) is source image pixels, and g (i, j) is output image pixel.I, j represent that pixel is arranged positioned at the i-th row jth.
Whether there is blink in step 4, the image by the multiple users of blink detection algorithm detection, deposit when detecting
In blink, it is determined that the image of the user is from live body.
The image of the user of each frame is converted into bianry image, the searching component pair in bianry image, according to setting in advance
The restrictive condition of the width of fixed eye element, height, horizontal range and vertical range, to the component that searches out to carrying out
Filter, acquisition meets the eye element pair of filter condition, and obtains region of search of the eye element to place.
According to the component for meeting filter condition to building eye template, calculated using normalizated correlation coefficient algorithm
Similar value between the eye template and the region of search, the region of search before and after continuous in the image of two framed users with
Judgment threshold of the difference between similar value between the eye template more than setting, it is determined that wrap in the image of the user
Containing blink, the image of the user is from live body.
A kind of schematic diagram of live body blink detection provided in an embodiment of the present invention is as shown in figure 3, concrete processing procedure includes：
Pretreatment：First image gray processing, then creates the difference image of each frame and former frame, resets threshold value, finally
The moving region of before and after two field pictures will be shown by bianry image.Then by the opening operation of morphological image computing, make
It is smooth that the profile of object becomes, the thinner protrusion of the narrow interruption of disconnection and elimination, so as to eliminate noise and photoenvironment
Impact.So applicability is also fine in the case where light is weaker.
Search eyes：When user people blinks, will flash a pair of component (binary maps of a pair of eyes in bianry image
As profile), after filtering out inappropriate component, then retrieved by profile, when it is determined that package count is 2 (eyes), just sentence
Break as eyes.Filtration is wide, the height according to two components of each centering, and the horizontal and vertical distance at two component centers.
Follow the trail of eyes：When find the component of the eyes for representing user to after, that bigger group is selected in this is to component
Part is used for creating eye template.The size of eye template is that the size for choosing component with that is directly proportional, and selects bigger that
One, because there is more monochrome informations, this will cause the associated score followed the trail of and obtain more accurate.Here bigger component is
The length, width and height of finger assembly are bigger.Because under low light environment, when user blinks, the eye contour shown on two-value difference image
(component to), typically small one and large one, that bigger building component eye template of selected shape.
How drawing template establishment：After selecting big component, (this border is just this component pair to obtain the boundary position of component
This eye answered are included), it is eye opening template to select the gray level image in this border.
Realize following the trail of using normalizated correlation coefficient, each two field picture is to using equation below：
F (x, y) is brightness of the frame of video in point (x, y),It is the mean value of current search area video frame brightness, t
(x, y) is brightness of the template picture in point (x, y),It is the mean value of template picture brightness.The result that above-mentioned formula 1 is calculated
The relevance scores between -1 and 1, it show eye template and screen frame search region a little between similitude.Point
Number shows low-level similitude closer to 0, and fraction represents that this is that it may matching eye opening template closer to 1.Using this
The key benefit that similarity measurement is followed the trail of is insensitive to being continually changing for ambient lighting conditions.When very dark,
Eye tracks and blink detection can work well.
Due to this method require it is a large amount of calculate, so region of search is limited to one around eyes of user very
Little region, greatly reducing needs the degree of correlation inquiry carried out in each frame to calculate, and allows system held stationary to run.
The matching process computing formula of eye template is：
Wherein, I represents image, and T represents eye template, and R represents result, eye template and image overlapping region x'=
0..w-1, sue for peace between y'=0..h-1.
The implication of the result of formula 2：Using the phase of template matches function successively calculation template and the overlapping region of picture to be measured
Like spend, and by result be stored in mapping graph as R in the middle of, that is to say, that the value of each point (x, y) in R images represents once phase
Like degree comparative result.
The relation of formula 1 and 2：Result R (x, y) of formula 2 is exactly the t (x, y) in the t (x, y) in formula 1, i.e. formula 1
It is calculated by formula 2.
After positioning eyes, when the eyes of user are when the process of blink is closed state, it is reduced with the similitude of human eye template.
Likewise, when blink is completed and the eyes of user are opened completely again, it obtains again similitude.When this similitude
One one liter of drop is judged as blink more than threshold value.It is in the display effect of bianry image, when user blinks, in the search of positioning
Inframe can flash the round dot of a white, and by flashing for this round dot blink is determined whether.
In the environment of using infrared light supply, because spectrally ultrared wavelength is more than luminous ray, for ripple
The long radiation in below 400nm or more than 700nm, human eye is helpless.But infrared light can but recognize face.So
On whole electromagnetic wave band, not only visible ray can be used for recognition of face, and infrared light can also.So the recognition of face for carrying out afterwards,
Matching also may be by infrared light supply to be carried out.
In sum, the face identification method based on blink detection provided in an embodiment of the present invention can be very by detection blink
Good differentiation detection is photo, sculpture or live body.Requirement of the blink detection method compared to ASM+Canny to light is very
High shortcoming (under the low light level, ASM is to human eye rim detection effect and bad to the locating effect and Canny of human eye), and it is of the invention
Difference processing is carried out to image using morphology, Dim light measurement condition can be well adapted to.
Compared in the not high shortcoming of different photoenvironment human face matching degrees.The present invention is varied multiple times using point operator
The brightness and contrast of registered human face photo, then carry out face matching checking, so as to the rate that improves that the match is successful.In light bet
The human face photo of volume, also can the match is successful under the low light level.
One of ordinary skill in the art will appreciate that：Accompanying drawing is the schematic diagram of one embodiment, module in accompanying drawing or
Flow process is not necessarily implemented necessary to the present invention.
As seen through the above description of the embodiments, those skilled in the art can be understood that the present invention can
Realize by the mode of software plus required general hardware platform.Based on such understanding, technical scheme essence
On prior art is contributed part in other words can be embodied in the form of software product, the computer software product
Can be stored in storage medium, such as ROM/RAM, magnetic disc, CD, including some instructions are used so that a computer equipment
(can be personal computer, server, either network equipment etc.) performs some of each embodiment of the invention or embodiment
Method described in part.
Each embodiment in this specification is described by the way of progressive, identical similar portion between each embodiment
Divide mutually referring to what each embodiment was stressed is the difference with other embodiment.Especially for device or
For system embodiment, because it is substantially similar to embodiment of the method, so describing fairly simple, related part is referring to method
The part explanation of embodiment.Apparatus and system embodiment described above is only schematic, wherein the conduct
Separating component explanation unit can be or may not be it is physically separate, can be as the part that unit shows or
Person may not be physical location, you can be located at a place, or can also be distributed on multiple NEs.Can be with
Select some or all of module therein according to the actual needs to realize the purpose of this embodiment scheme.The common skill in this area
Art personnel are not in the case where creative work is paid, you can to understand and implement.
The above, the only present invention preferably specific embodiment, but protection scope of the present invention is not limited thereto,
Any those familiar with the art the invention discloses technical scope in, the change or replacement that can be readily occurred in,
All should be included within the scope of the present invention.Therefore, protection scope of the present invention should be with scope of the claims
It is defined.
Claims (6)
1. a kind of face identification method based on blink detection, it is characterised in that include：
The image of user is shot by camera, facial image is detected from the image of the user by human-face detector；
The facial image is matched with the chartered facial image for prestoring by face matching algorithm, is worked as institute
The result for stating matching is similarity more than during the matching threshold for setting, it is determined that the facial image has been registered；
Detected by blink detection algorithm in the image of multiple users with the presence or absence of blink, there is blink when detecting, then
Determine the image of the user from live body.
2. method according to claim 1, it is characterised in that the described image that user is shot by camera, is passed through
Human-face detector detects facial image from the image of the user, including：
Using Adaboost method construct human-face detectors, the human-face detector is the cascade classifier of screening type, cascade
Each node arranges that arrange judgment threshold on each node, position node more rearward is comprising more according to cascade sort algorithm
Many Weak Classifiers；
The image of the user of each frame that camera is shot is input into the human-face detector, and the human-face detector is successively
The image of the user of each frame is detected, from the image of the user facial image is detected.
3. according to claim 1 or claim 2, it is characterised in that it is described by face matching algorithm by the people
Face image is matched with the chartered facial image for prestoring, when the result of the matching is that similarity is more than setting
Matching threshold when, it is determined that the facial image has been registered, including：
After facial image is detected in the image from the user of present frame, single channel will be converted into after the facial image cutting
Image, by the registered facial image for prestoring single channel image is converted into, and described two single channel images are made into 4 kinds directly
Square figure compares, and draws 4 Similarity values：CV_COMP_CHISQR card sides, CV_COMP_BHATTACHARYYA distances, CV_
The COMP_CORREL degrees of correlation, CV_COMP_INTERSECT intersects coefficient；
Final similarity between described two single channel images is determined according to 4 Similarity values, when described final similar
Degree is more than similarity threshold set in advance, then judge that the match is successful for described two single channel images, determines the facial image
Register.
4. method according to claim 3, it is characterised in that described method also includes：
When the final similarity is not more than similarity threshold set in advance, then judge that described two single channel image matchings are lost
Lose, the brightness and/or contrast to the registered facial image is adjusted, regenerate the facial image after adjustment
Single channel image；
Two single channel images are matched again according to above-mentioned matching process, when two single channel images, the match is successful, determines institute
State facial image to have registered；When two single channel images, it fails to match, then continue to the bright of the registered facial image
Degree and/or contrast are adjusted, and continuation again matches two single channel images according to above-mentioned matching process, when it fails to match
Number of times more than setting numerical value, it is determined that the facial image is not registered.
5. method according to claim 4, it is characterised in that described method also includes：
Brightness and/or contrast to the registered facial image is adjusted using an Operators Algorithm, and formula is as follows：
G (i, j)=α f (i, j)+β
Parameter alpha>0, it is gain, for adjusting contrast, β is offset parameter, and for adjusting brightness, f (i, j) is source images picture
Element, g (i, j) is output image pixel, and i, j represents that pixel is arranged positioned at the i-th row jth.
6. method according to claim 4, it is characterised in that described detects multiple use by blink detection algorithm
With the presence or absence of blink in the image at family, there is blink when detecting, it is determined that the image of the user from live body, including：
The image of the user of each frame is converted into bianry image, the searching component pair in the bianry image, according to pre-
The restrictive condition of the width of the eye element for first setting, height, horizontal range and vertical range, to the component that searches out to carrying out
Filter, acquisition meets the eye element pair of filter condition, and obtains region of search of the eye element to place；
According to the component for meeting filter condition to building eye template, calculated using normalizated correlation coefficient algorithm described
Similar value between eye template and the region of search, the region of search before and after continuous in the image of two framed users with it is described
Judgment threshold of the difference between similar value between eye template more than setting, it is determined that include in the image of the user and blink
Eye, the image of the user is from live body.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611039329.4A CN106682578B (en) | 2016-11-21 | 2016-11-21 | Weak light face recognition method based on blink detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611039329.4A CN106682578B (en) | 2016-11-21 | 2016-11-21 | Weak light face recognition method based on blink detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106682578A true CN106682578A (en) | 2017-05-17 |
| CN106682578B CN106682578B (en) | 2020-05-05 |
Family
ID=58867269
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611039329.4A Active CN106682578B (en) | 2016-11-21 | 2016-11-21 | Weak light face recognition method based on blink detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106682578B (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107730364A (en) * | 2017-10-31 | 2018-02-23 | 北京麒麟合盛网络技术有限公司 | user identification method and device |
| CN107820591A (en) * | 2017-06-12 | 2018-03-20 | 美的集团股份有限公司 | Control method, controller, Intelligent mirror and computer-readable recording medium |
| CN107844764A (en) * | 2017-10-31 | 2018-03-27 | 广东欧珀移动通信有限公司 | Image processing method, device, electronic equipment and computer-readable recording medium |
| CN108133177A (en) * | 2017-12-06 | 2018-06-08 | 山东超越数控电子股份有限公司 | A kind of method for improving Face datection reliability |
| CN108229376A (en) * | 2017-12-29 | 2018-06-29 | 百度在线网络技术（北京）有限公司 | For detecting the method and device of blink |
| CN108363999A (en) * | 2018-03-22 | 2018-08-03 | 百度在线网络技术（北京）有限公司 | Operation based on recognition of face executes method and apparatus |
| CN108985328A (en) * | 2018-06-08 | 2018-12-11 | 佛山市顺德区中山大学研究院 | For differentiating the construction method and its system of the deep learning model of corneal ulceration |
| CN109658533A (en) * | 2018-11-23 | 2019-04-19 | 深圳市沃特沃德股份有限公司 | Method of registering, system and the intelligent terminal of video conference |
| CN109858375A (en) * | 2018-12-29 | 2019-06-07 | 深圳市软数科技有限公司 | Living body faces detection method, terminal and computer readable storage medium |
| CN110334637A (en) * | 2019-06-28 | 2019-10-15 | 百度在线网络技术（北京）有限公司 | Human face in-vivo detection method, device and storage medium |
| CN113331824A (en) * | 2021-04-20 | 2021-09-03 | 北京九辰智能医疗设备有限公司 | Blink judgment method and device based on laser measurement |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101702198A (en) * | 2009-11-19 | 2010-05-05 | 浙江大学 | Identification method for video and living body faces based on background comparison |
| CN101770613A (en) * | 2010-01-19 | 2010-07-07 | 北京智慧眼科技发展有限公司 | Social insurance identity authentication method based on face recognition and living body detection |
| CN102622588A (en) * | 2012-03-08 | 2012-08-01 | 无锡数字奥森科技有限公司 | Dual-certification face anti-counterfeit method and device |
| CN102789572A (en) * | 2012-06-26 | 2012-11-21 | 五邑大学 | Living body face safety certification device and living body face safety certification method |
| CN103400122A (en) * | 2013-08-20 | 2013-11-20 | 江苏慧视软件科技有限公司 | Method for recognizing faces of living bodies rapidly |
| CN103434484A (en) * | 2013-08-20 | 2013-12-11 | 安科智慧城市技术（中国）有限公司 | Vehicle-mounted identification and authentication device, mobile terminal and intelligent vehicle key control system and method |
| CN103593598A (en) * | 2013-11-25 | 2014-02-19 | 上海骏聿数码科技有限公司 | User online authentication method and system based on living body detection and face recognition |
| CN103886301A (en) * | 2014-03-28 | 2014-06-25 | 中国科学院自动化研究所 | Human face living detection method |
| CN104361326A (en) * | 2014-11-18 | 2015-02-18 | 新开普电子股份有限公司 | Method for distinguishing living human face |
| CN105243386A (en) * | 2014-07-10 | 2016-01-13 | 汉王科技股份有限公司 | Face living judgment method and system |
| CN105993022A (en) * | 2016-02-17 | 2016-10-05 | 香港应用科技研究院有限公司 | Recognition and authentication method and system using facial expression |
-
2016
- 2016-11-21 CN CN201611039329.4A patent/CN106682578B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101702198A (en) * | 2009-11-19 | 2010-05-05 | 浙江大学 | Identification method for video and living body faces based on background comparison |
| CN101770613A (en) * | 2010-01-19 | 2010-07-07 | 北京智慧眼科技发展有限公司 | Social insurance identity authentication method based on face recognition and living body detection |
| CN102622588A (en) * | 2012-03-08 | 2012-08-01 | 无锡数字奥森科技有限公司 | Dual-certification face anti-counterfeit method and device |
| CN102789572A (en) * | 2012-06-26 | 2012-11-21 | 五邑大学 | Living body face safety certification device and living body face safety certification method |
| CN103400122A (en) * | 2013-08-20 | 2013-11-20 | 江苏慧视软件科技有限公司 | Method for recognizing faces of living bodies rapidly |
| CN103434484A (en) * | 2013-08-20 | 2013-12-11 | 安科智慧城市技术（中国）有限公司 | Vehicle-mounted identification and authentication device, mobile terminal and intelligent vehicle key control system and method |
| CN103593598A (en) * | 2013-11-25 | 2014-02-19 | 上海骏聿数码科技有限公司 | User online authentication method and system based on living body detection and face recognition |
| CN103886301A (en) * | 2014-03-28 | 2014-06-25 | 中国科学院自动化研究所 | Human face living detection method |
| CN105243386A (en) * | 2014-07-10 | 2016-01-13 | 汉王科技股份有限公司 | Face living judgment method and system |
| CN104361326A (en) * | 2014-11-18 | 2015-02-18 | 新开普电子股份有限公司 | Method for distinguishing living human face |
| CN105993022A (en) * | 2016-02-17 | 2016-10-05 | 香港应用科技研究院有限公司 | Recognition and authentication method and system using facial expression |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107820591A (en) * | 2017-06-12 | 2018-03-20 | 美的集团股份有限公司 | Control method, controller, Intelligent mirror and computer-readable recording medium |
| CN107844764A (en) * | 2017-10-31 | 2018-03-27 | 广东欧珀移动通信有限公司 | Image processing method, device, electronic equipment and computer-readable recording medium |
| CN107730364A (en) * | 2017-10-31 | 2018-02-23 | 北京麒麟合盛网络技术有限公司 | user identification method and device |
| CN108133177A (en) * | 2017-12-06 | 2018-06-08 | 山东超越数控电子股份有限公司 | A kind of method for improving Face datection reliability |
| CN108229376B (en) * | 2017-12-29 | 2022-06-03 | 百度在线网络技术（北京）有限公司 | Method and device for detecting blinking |
| CN108229376A (en) * | 2017-12-29 | 2018-06-29 | 百度在线网络技术（北京）有限公司 | For detecting the method and device of blink |
| CN108363999A (en) * | 2018-03-22 | 2018-08-03 | 百度在线网络技术（北京）有限公司 | Operation based on recognition of face executes method and apparatus |
| CN108985328A (en) * | 2018-06-08 | 2018-12-11 | 佛山市顺德区中山大学研究院 | For differentiating the construction method and its system of the deep learning model of corneal ulceration |
| CN109658533A (en) * | 2018-11-23 | 2019-04-19 | 深圳市沃特沃德股份有限公司 | Method of registering, system and the intelligent terminal of video conference |
| CN109858375A (en) * | 2018-12-29 | 2019-06-07 | 深圳市软数科技有限公司 | Living body faces detection method, terminal and computer readable storage medium |
| CN109858375B (en) * | 2018-12-29 | 2023-09-26 | 简图创智(深圳)科技有限公司 | Living body face detection method, terminal and computer readable storage medium |
| CN110334637A (en) * | 2019-06-28 | 2019-10-15 | 百度在线网络技术（北京）有限公司 | Human face in-vivo detection method, device and storage medium |
| CN113331824A (en) * | 2021-04-20 | 2021-09-03 | 北京九辰智能医疗设备有限公司 | Blink judgment method and device based on laser measurement |
| CN113331824B (en) * | 2021-04-20 | 2023-09-26 | 北京九辰智能医疗设备有限公司 | Blink judgment method and device based on laser measurement |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106682578B (en) | 2020-05-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106682578A (en) | Human face recognition method based on blink detection | |
| CN105868689B (en) | A kind of face occlusion detection method based on concatenated convolutional neural network | |
| D'Orazio et al. | An algorithm for real time eye detection in face images | |
| Chaudhari et al. | Face detection using viola jones algorithm and neural networks | |
| Proenca | Iris recognition: On the segmentation of degraded images acquired in the visible wavelength | |
| CN106228137A (en) | A kind of ATM abnormal human face detection based on key point location | |
| CN104036278B (en) | The extracting method of face algorithm standard rules face image | |
| CN106600631A (en) | Multiple target tracking-based passenger flow statistics method | |
| Reese et al. | A comparison of face detection algorithms in visible and thermal spectrums | |
| CN109299690B (en) | Method capable of improving video real-time face recognition precision | |
| Suma et al. | Real time face recognition of human faces by using LBPH and Viola Jones algorithm | |
| CN106650574A (en) | Face identification method based on PCANet | |
| CN110276320A (en) | Guard method, device, equipment and storage medium based on recognition of face | |
| Radu et al. | A robust sclera segmentation algorithm | |
| CN108171215B (en) | Face camouflage detection and camouflage type detection method based on low-rank variation dictionary and sparse representation classification | |
| CN113221655A (en) | Face spoofing detection method based on feature space constraint | |
| Lin et al. | Face detection based on skin color segmentation and neural network | |
| CN102831379B (en) | facial image recognition method and device | |
| D'orazio et al. | A neural system for eye detection in a driver vigilance application | |
| Hsiao et al. | EfficientNet based iris biometric recognition methods with pupil positioning by U-net | |
| CN107220623A (en) | A kind of face identification method and system | |
| Hassan et al. | SIPFormer: Segmentation of multiocular biometric traits with transformers | |
| Mary et al. | Human identification using periocular biometrics | |
| Das et al. | Human face detection in color images using HSV color histogram and WLD | |
| D'orazio et al. | Eye detection in face images for a driver vigilance system |
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 |