CN103093210B - Method and device for glasses identification in face identification - Google Patents
Method and device for glasses identification in face identification Download PDFInfo
- Publication number
- CN103093210B CN103093210B CN201310027751.8A CN201310027751A CN103093210B CN 103093210 B CN103093210 B CN 103093210B CN 201310027751 A CN201310027751 A CN 201310027751A CN 103093210 B CN103093210 B CN 103093210B
- Authority
- CN
- China
- Prior art keywords
- predeterminable area
- glasses
- threshold value
- predetermined threshold
- region
- 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.)
- Active
Links
Abstract
The invention discloses a method and a device for glasses identification in face identification. The method comprises the steps: obtaining a facial image, and clipping a first preset area in the obtained facial image, wherein a covered range of the first preset area comprises a human eye area and a part or all area covered by a pair of glasses; and judging types of the pair of the glasses according to the first preset area. According to the technical scheme of the method and the device for the glasses identification in the face identification, facial images with the pair of glasses and reflective can be accurately judged out. Through the method and the device for the glasses identification in the face identification, states of eyes in the human eye locating process of the face identification are classified, different methods are adopted to locate human eyes in different situations, and therefore speed and accuracy of human eye locating are improved.
Description
Technical field
The present invention relates to image procossing and area of pattern recognition, in particular to a kind of mirror of glasses in recognition of face
Other method and device.
Background technology
In epoch of developing rapidly of informationization now, particularly ecommerce and social safety field, how precise Identification
The identity of one people, protection information safety, have become as social problem urgently to be resolved hurrily.Identification authentication mode in correlation technique
Forge, lose and the disadvantage such as inconvenient to carry due to existing, being increasingly difficult to meet the demand of social development, therefore biological characteristic
Identification has obtained increasing attention.With the continuous development of the subjects such as computer science and pattern recognition, make this field
The cost of realizing of high-performance automatic identification technology is reduced to acceptable degree.And phase identification in face is all biological characteristics knowledges
One of technology being most widely used in other method.It is fast, square that face phase identification has speed with respect to other living things feature recognitions
Just, the advantages of, contactless, it is by the most natural, the most direct means of identity validation.
Face phase identification is mainly by Face datection, human eye positioning, facial characteristics registration, Feature extraction and recognition four part structure
Become.In the identification of face phase, no matter being the global characteristics using face or local feature, the change of face orientation and size is right
Identification has significant impact it is therefore desirable to carry out registration to face picture.And eyes are as special in human face
Levy, there is compared with the organs such as nose, face more abundant information, two spacing are subject to the shadow of illumination or expression shape change simultaneously
Sound is less, and the straight line that two eye pupil holes are located deflects with the deflection of face, and therefore human eye can be used as the standard point of face registration.
More improve with the requirement to various aspects such as the accuracy and speeds of eyes location algorithm for the face phase identifying system.Therefore, how
Fast and effectively eyes are positioned, become one of face phase identifying system important study portion.People a lot of at present
Will wearing spectacles, wear sunglasses and will lead to not carry out recognition of face, regardless of whether wearing black surround glasses or common
Glasses all can produce on human eye positioning be affected to some extent, and the direct projection of under the natural light or infrared light light of wearing glasses is all
Can cause glasses reflection, this reflective sometimes even the whole region of human eye all can be sheltered from.Therefore, in recognition of face
Cheng Zhong, needs the glasses worn are differentiated.In correlation technique, directly pass through to judge whether glasses are reflective thus carrying out human eye
The technical scheme speed of positioning is slower, and accuracy is poor.
Content of the invention
The invention provides the discrimination method of glasses and device in a kind of recognition of face, to solve to lack in correlation technique
In the technical scheme judging before whether glasses are reflective, the type of human eye wearing spectacles to be differentiated in face recognition process
Problem.
According to an aspect of the invention, it is provided in a kind of recognition of face glasses discrimination method.
Included according to the discrimination method of glasses in the recognition of face of the present invention:Obtain facial image, and in the people getting
The first predeterminable area is intercepted, wherein, the coverage of the first predeterminable area includes in face image:Human eye area and glasses cover
Some or all of region;Judge the type of glasses according to the first predeterminable area.
Preferably, included according to the type that the first predeterminable area judges glasses:Statistics gray value in the first predeterminable area
Number less than the pixel of the first predetermined threshold value;Number according to the pixel less than the first predetermined threshold value and the first preset areas
The ratio of the number of the pixel in domain and the comparative result of the second predetermined threshold value, judge whether the type of glasses is sunglasses.
Preferably, included according to the type that the first predeterminable area judges glasses:Homomorphic filtering is carried out to the first predeterminable area
Process;The disposal of gentle filter is carried out to the first predeterminable area after homomorphic filtering is processed;To after the disposal of gentle filter
The first predeterminable area carry out neighborhood minimum Filtering Processing;To the first predeterminable area after neighborhood minimum Filtering Processing
It is fixed ratio binary conversion treatment;The first predeterminable area after fixed proportion binary conversion treatment intercepts second preset
Region, wherein, the second predeterminable area is symmetrical with regard to face axis, and the second predeterminable area includes:Left eye picture frame and right eye
Connector between picture frame;Ratio according to connector region proportion and the 3rd predetermined threshold value in the second predeterminable area
Relatively result, judges whether the type of glasses is black surround glasses.
Preferably, included according to the type that the first predeterminable area judges glasses:Sobel operator is carried out to the first predeterminable area
Filtering Processing;Binary conversion treatment is carried out to the first predeterminable area after Sobel operator filtering is processed;To at binaryzation
The first predeterminable area after reason carries out Morphological scale-space, extracts one or more connected regions;To one or more connected regions
It is integrated projection to calculate, judge that whether the type of glasses is the other kinds of glasses in addition to sunglasses or black surround glasses.
Preferably, judge glasses type be black surround glasses or other classes in addition to sunglasses or black surround glasses
After the glasses of type, judge glasses whether reflective inclusion under light irradiation:3rd preset areas are intercepted on the first predeterminable area
Domain, wherein, the coverage of the 3rd predeterminable area includes:Human eye area, and the 3rd predeterminable area is less than the first predeterminable area;
Left eye region and right eye region is determined respectively in the 3rd predeterminable area;Statistics left eye region and right eye respectively
In region, gray value is more than the number of the pixel of the 4th predetermined threshold value;Choose the 5th predetermined threshold value N from 0 to 255, then
Start to 255 to terminate to choose each positive integer successively from N, calculate gray value in left eye region and right eye region respectively
It is equal to the product of each number of the pixel of positive integer chosen and this positive integer, and calculate the summation of whole result of product,
Then using the summation calculating divided by the area of left eye region and right eye region, place is normalized to summation
Reason, wherein, N is 0 or positive integer and N≤255;Individual more than the pixel of the 4th predetermined threshold value according to the gray value counting
Whether several normalization results whether being more than the 6th predetermined threshold value and calculating are more than the 7th predetermined threshold value, judge described face
Whether the glasses in image are reflective.
According to a further aspect in the invention, there is provided the identification device of glasses in a kind of recognition of face.
Included according to the identification device of glasses in the recognition of face of the present invention:Acquisition module, for obtaining facial image, and
Intercept the first predeterminable area in the facial image getting, wherein, the coverage of the first predeterminable area includes:Human eye area
And some or all of region that glasses cover;Judge module, for judging the type of glasses according to the first predeterminable area.
Preferably, judge module includes:First statistic unit, for statistics, in the first predeterminable area, gray value is less than the
The number of the pixel of one predetermined threshold value;First judging unit, for the number according to the pixel less than the first predetermined threshold value
With the ratio of the number of the pixel in the first predeterminable area and the comparative result of the second predetermined threshold value, judge that the type of glasses is
No for sunglasses.
Preferably, judge module includes:First processing units, for carrying out homomorphic filtering process to the first predeterminable area;
Second processing unit, for carrying out the disposal of gentle filter to the first predeterminable area after homomorphic filtering is processed;3rd process
Unit, for carrying out neighborhood minimum Filtering Processing to the first predeterminable area after the disposal of gentle filter;Fourth process list
Unit, for being fixed ratio binary conversion treatment to the first predeterminable area after neighborhood minimum Filtering Processing;First section
Take unit, for intercepting the second predeterminable area on the first predeterminable area after fixed proportion binary conversion treatment, wherein, the
Two predeterminable areas are symmetrical with regard to face axis, and the second predeterminable area includes:Company between left eye picture frame and right eye picture frame
Fitting;Second judging unit, for according to connector region in the second predeterminable area proportion and the 3rd default threshold
The comparative result of value, judges whether the type of glasses is black surround glasses.
Preferably, judge module includes:5th processing unit, for carrying out Sobel operator filtering to the first predeterminable area
Process;6th processing unit, for carrying out binary conversion treatment to the first predeterminable area after Sobel operator filtering is processed;
Extraction unit, for the first predeterminable area after binary conversion treatment is carried out with Morphological scale-space, extracts one or more companies
Logical region;3rd judging unit, calculates for one or more connected regions are integrated with projection, judges that the type of glasses is
No is other kinds of glasses in addition to sunglasses or black surround glasses.
Preferably, judge module, is additionally operable to the type judging glasses and for black surround glasses or removes sunglasses or black surround
After other kinds of glasses outside glasses, judge whether the glasses in described facial image are reflective, and judge module also includes:
Second interception unit, for intercepting the 3rd predeterminable area, wherein, the coverage of the 3rd predeterminable area on the first predeterminable area
Including:Human eye area, and the 3rd predeterminable area is less than the first predeterminable area;Determining unit, in the 3rd predeterminable area
Determine left eye region and right eye region respectively;Second statistic unit, for counting left eye region and the right side respectively
In eye region, gray value is more than the number of the pixel of the 4th predetermined threshold value;3rd statistic unit, for from 0 to 255
Choose the 5th predetermined threshold value N, then from N start to 255 terminate choose each positive integer successively, respectively calculate left eye region and
In right eye region, gray value is equal to the product of each number of the pixel of positive integer chosen and this positive integer, and calculates
All summations of result of product, then using the summation calculating divided by left eye region and right eye region area,
Summation is normalized, wherein, N is 0 or positive integer and N≤255;4th judging unit, counts for basis
Gray value be more than the 4th predetermined threshold value pixel number whether more than the 6th predetermined threshold value and the normalization that calculates
Whether result is more than the 7th predetermined threshold value, judges whether the glasses in described facial image are reflective.
By the present invention, using acquisition facial image, and intercept the first predeterminable area in the facial image getting, its
In, the coverage of the first predeterminable area includes:Human eye area and some or all of region of glasses covering;Judge glasses
Type and/or glasses light irradiation under whether reflective, solve to lack in correlation technique and sentencing in face recognition process
To the type of human eye wearing spectacles and asking of the whether reflective technical scheme being differentiated of glasses before whether disconnected glasses are reflective
Topic, and then can quickly and accurately judge wearing spectacles and reflective facial image.Human eye position fixing process to recognition of face
The state of middle eyes is classified, and the human eye for different situations is positioned using different methods, thus improve human eye
The speed of positioning and accuracy.
Brief description
Accompanying drawing described herein is used for providing a further understanding of the present invention, constitutes the part of the application, this
Bright schematic description and description is used for explaining the present invention, does not constitute inappropriate limitation of the present invention.In the accompanying drawings:
Fig. 1 is the flow chart of the discrimination method of glasses in recognition of face according to embodiments of the present invention;
Fig. 2 is the schematic diagram of the first untreated predeterminable area according to the preferred embodiment of the present invention one;
Fig. 3 is the schematic diagram of the first predeterminable area being processed through homomorphic filtering according to the preferred embodiment of the present invention one;
Fig. 4 is the showing of the first predeterminable area according to the preferred embodiment of the present invention one through neighborhood minimum Filtering Processing
It is intended to;
Fig. 5 is in the first predeterminable area pair through fixed proportion binary conversion treatment according to the preferred embodiment of the present invention one
Compare schematic diagram;
Fig. 6 a is the schematic diagram of the first untreated predeterminable area according to the preferred embodiment of the present invention two;
Fig. 6 b is the schematic diagram of the first untreated predeterminable area according to the preferred embodiment of the present invention three;
Fig. 7 a is the schematic diagram that according to the preferred embodiment of the present invention two, Fig. 6 a is carried out with Sobel operator filtering process;
Fig. 7 b is the schematic diagram that according to the preferred embodiment of the present invention three, Fig. 6 b is carried out with Sobel operator filtering process;
Fig. 8 a is the schematic diagram that according to the preferred embodiment of the present invention two, Fig. 7 a is carried out with binary conversion treatment;
Fig. 8 b is the schematic diagram that according to the preferred embodiment of the present invention three, Fig. 7 b is carried out with binary conversion treatment;
Fig. 9 a is the schematic diagram that according to the preferred embodiment of the present invention two, Fig. 8 a is carried out with Morphological scale-space;
Fig. 9 b is the schematic diagram that according to the preferred embodiment of the present invention three, Fig. 8 b is carried out with Morphological scale-space;
Figure 10 is the schematic diagram of the 3rd untreated predeterminable area according to the preferred embodiment of the invention;
Figure 11 is the judgement side of glasses and glasses reflection in face recognition process according to the preferred embodiment of the invention
The flow chart of method;
Figure 12 is the structured flowchart of the identification device of glasses in recognition of face according to embodiments of the present invention;
Figure 13 is the structured flowchart of the identification device of glasses in recognition of face according to the preferred embodiment of the invention.
Specific embodiment
To describe the present invention in detail below with reference to accompanying drawing and in conjunction with the embodiments.It should be noted that not conflicting
In the case of, the embodiment in the application and the feature in embodiment can be mutually combined.
Fig. 1 is the flow chart of the discrimination method of glasses in recognition of face according to embodiments of the present invention.As shown in figure 1, should
Method can include following process step:
Step S102:Obtain facial image, and intercept the first predeterminable area in the facial image getting, wherein, the
The coverage of one predeterminable area includes:Human eye area and some or all of region of glasses covering;
Step S104:Judge the type of described glasses according to the first predeterminable area.
In correlation technique, lack in face recognition process judge glasses before whether reflective to human eye wearing spectacles
The technical scheme that type is differentiated.Using method as shown in Figure 1, using acquisition facial image, and in the face getting
The first predeterminable area is intercepted, wherein, the coverage of the first predeterminable area includes in image:Human eye area and glasses cover
Some or all of region;Judge the type of the glasses in described facial image according to the first predeterminable area, solve related skill
Lack in art and judging before whether glasses are reflective, the type of human eye wearing spectacles to be differentiated in face recognition process
The problem of technical scheme, and then can quickly and accurately judge wearing spectacles and further determine that reflective facial image, right
In the human eye position fixing process of recognition of face, the state of eyes is classified, and the human eye for different situations is entered using different methods
Row positioning, thus improve speed and the accuracy of human eye positioning.
In a preferred embodiment, Fig. 2 is the first untreated predeterminable area according to the preferred embodiment of the present invention one
Schematic diagram.As shown in Fig. 2 intercept the first predeterminable area in the facial image getting can hang down using with face axis
Be parallel to each other two straight straight lines are set in above brows and below the bridge of the nose respectively, and the region between two straight lines is as above-mentioned
First predeterminable area.The size of hypothesis facial image is M × N, then the size of the first predeterminable area is then M' × N ', and M ' is permissible
TakeWithBetween part, and N' can take the part between 1 to N.The intercepting of certain first predeterminable area is acceptable
Using other modes, as long as some or all of region that human eye area and glasses cover can be included, no longer superfluous herein
State.
Preferably, in step S104, following operation can be included according to the type that the first predeterminable area judges glasses:
Step S1:Statistics gray value in the first predeterminable area is less than the number of the pixel of the first predetermined threshold value;
Gray scale refers to the color depth at black white image midpoint, and typically from 0 to 255, white is 255 to scope, and black is 0, therefore
Black and white picture is also referred to as gray level image, has a wide range of applications in medical science, field of image recognition.
Step S2:Pixel in number according to the pixel less than the first predetermined threshold value and the first predeterminable area
The ratio of number and the comparative result of the second predetermined threshold value, judge whether the type of glasses is sunglasses.
In a preferred embodiment, in number and the ocular being covered by glasses of the pixel less than the first predetermined threshold value
The ratio of the number of interior pixel be more than the second predetermined threshold value when, you can judge glasses type be sunglasses.
In a preferred embodiment, after the type judging glasses is sunglasses, because sunglasses can lead to not into pedestrian
Face identifies, now directly will export judged result, and need not carry out judging the follow-up process such as whether glasses reflective again.
Preferably, in step S104, be may comprise steps of according to the type that the first predeterminable area judges glasses:
Step S3:Homomorphic filtering process is carried out to the first predeterminable area;
In a preferred embodiment, homomorphic filtering is a kind of image processing method that frequency filter and greyscale transformation combine
Method, it relies on the basis that the illumination of image or Reflectivity Model are processed as frequency domain, right using compression brightness range and enhancing
To improve the quality of image than degree.Image procossing can be made to meet the non-linear spy for luminosity response for the human eye using the method
Property, thus avoiding directly carrying out, to image, the distortion that Fourier transform processing is led to.
Fig. 3 is the schematic diagram of the first predeterminable area being processed through homomorphic filtering according to the preferred embodiment of the present invention one.
The gray scale of image f (x, y) can be analyzed to:f(x,y)=i(x,y)r(x,y)
In the preferred embodiment, in above formula, f (x, y) represents the image of the first predeterminable area, wherein, the property of i (x, y)
Depending on irradiation source, it is grading function, and r (x, y) depends on the characteristic of imaging object, is reflective function.
Specific calculation procedure is as follows:
(1)Multiplication is converted into addition, carries out natural logrithm(ln)Computing:
z(x,y)=lnf(x,y)=lni(x,y)+lnr(x,y);
Z (x, y) represents the image that above-mentioned f (x, y) obtains after logarithm operation;
(2)Image is transformed in frequency domain, carries out Fourier transform:
Z(u,v)=I(u,v)+R(u,v);
Z (u, v) represents that above-mentioned z (x, y), through Fourier transformation, is changed to the image obtaining after frequency domain by transform of spatial domain;
(3)Press irradiation component, spread reflection component, carry out homomorphism and increase clear computing:
H(u,v)Z(u,v)=H(u,v)I(u,v)+H(u,v)R(u,v);
Wherein, H (u, v) is homomorphic filter, similar with the primitive form of ideal highpass filter, can be according to difference
Picture characteristics and needs, from different H (u, v), H (u, v) Z (u, v) represent above-mentioned Z (u, v) through homomorphism increase clear computing
The image obtaining afterwards;
(4)Carry out inverse fourier transform, obtain corresponding spatial domain representation
I'(x, y)=ifft [H (u, v) I (u, v)];
R'(x, y)=ifft [H (u, v) R (u, v)];
(5)Carry out exp computing, obtain final result:
f'(x,y)=i'(x,y)r'(x,y);
Wherein, i'(x, y) and r'(x, y) is the grading function regenerating and reflective function;F ' (x, y) is through homomorphism
The image generating after Filtering Processing.Detail contrast is poor, differentiate unclear image, after processing, image using homomorphic filtering
Brightness ratio is more uniform, and details can strengthen.
Step S4:The disposal of gentle filter is carried out to the first predeterminable area after homomorphic filtering is processed;
In a preferred embodiment, using Gaussian filter, the image of the first predeterminable area through homomorphic filtering can be entered
Row the disposal of gentle filter.Gaussian filter is shape according to the Gaussian function linear smoothing filter to select weights.Gauss
Smoothing filter has good effect to the noise removing Normal Distribution.The expression formula of Gaussian filter is as follows:
Step S5:Neighborhood minimum Filtering Processing is carried out to the first predeterminable area after the disposal of gentle filter;
In a preferred embodiment, Fig. 4 is through neighborhood minimum Filtering Processing according to the preferred embodiment of the present invention one
The schematic diagram of the first predeterminable area.Neighborhood minimum filtering is that each pixel in image is traveled through, and value herein will be by this
Point certain neighborhood window in institute a little in minimum pixel value replacement.As shown in figure 4, after adopting neighborhood minimum Filtering Processing,
The region of the glasses in image and eyes more highlights.Following formula is neighborhood minimum Filtering Formula.
Wherein, I (x, y) is the pixel value of point (x, y) in the first predeterminable area of gaussian filtering, I'(x, y) is point
Pixel value after (x, y) mini-value filtering.
Step S6:The first predeterminable area after neighborhood minimum Filtering Processing is fixed at ratio binaryzation
Reason;
Fixed proportion binaryzation refers to determine that the gray value from 0 beginning accounts for the proportion of total gray value, using statistics with histogram
Corresponding gray value when being met setting proportion, that is, obtain the threshold value of segmentation figure picture.
Step S7:The first predeterminable area after fixed proportion binary conversion treatment intercepts the second predeterminable area, its
In, the second predeterminable area is symmetrical with regard to face axis, and the second predeterminable area includes:Between left eye picture frame and right eye picture frame
Connector;
Step S8:Comparison according to connector region proportion and the 3rd predetermined threshold value in the second predeterminable area
As a result, whether the type judging glasses is black surround glasses.
In a preferred embodiment, Fig. 5 is through fixed proportion binary conversion treatment according to the preferred embodiment of the present invention one
The first predeterminable area contrast schematic diagram.The company of middle black is can be clearly seen that in the latter half two image of Fig. 5
Fitting region, the mid portion vertically intercepting the first area after binaryzation is as the second predeterminable area(Can be first preset
The longitudinal central axis line in region extends the region of certain pixel respectively to both sides), when connector region is in the second predeterminable area
When interior proportion is more than three predetermined threshold value, you can the type judging glasses is black surround glasses.
Preferably, in step S104, following operation can be included according to the type that the first predeterminable area judges glasses:
Step S9:First predeterminable area is carried out with Sobel operator filtering process;Fig. 6 a is according to the preferred embodiment of the present invention
The schematic diagram of two the first untreated predeterminable area.Fig. 6 b is untreated according to the preferred embodiment of the present invention three
The first predeterminable area schematic diagram.Fig. 7 a is to carry out Sobel operator filtering according to the preferred embodiment of the present invention two to Fig. 6 a
The schematic diagram processing.Fig. 7 b is the signal that according to the preferred embodiment of the present invention three, Fig. 6 b is carried out with Sobel operator filtering process
Figure.As shown in figs. 7 a and 7b.
Sobel operator refers in Edge check, a kind of conventional template operator.Sobel operator has two:One is inspection
Survey horizontal edge, another is detection vertical edges edge.Operator is as follows:
Step S10:Fig. 8 a is the schematic diagram that according to the preferred embodiment of the present invention two, Fig. 7 a is carried out with binary conversion treatment.Figure
8b is the schematic diagram that according to the preferred embodiment of the present invention three, Fig. 7 b is carried out with binary conversion treatment.As figures 8 a and 8 b show, right
The first predeterminable area after Sobel operator filtering is processed carries out binary conversion treatment;
Step S11:Fig. 9 a is the schematic diagram that according to the preferred embodiment of the present invention two, Fig. 8 a is carried out with Morphological scale-space.Figure
9b is the schematic diagram that according to the preferred embodiment of the present invention three, Fig. 8 b is carried out with Morphological scale-space.As shown in figures 9 a and 9b, right
The first predeterminable area after binary conversion treatment carries out Morphological scale-space, extracts one or more connected regions;
Step S12:One or more connected regions are integrated with projection calculate, judge whether the type of glasses is ink removing
Other kinds of glasses outside mirror or black surround glasses.
Integral projection refers to image is carried out with projection both horizontally and vertically, calculates the gray scale in both direction respectively
Value sum.
Preferably, in step S104, for black surround glasses or sunglasses or black surround eye are removed in the type judging glasses
After other kinds of glasses outside mirror, judge that whether reflective glasses under light irradiation and may comprise steps of:
Step S13:3rd predeterminable area, wherein, the coverage of the 3rd predeterminable area are intercepted on the first predeterminable area
Including:Human eye area, and the 3rd predeterminable area is less than the first predeterminable area;
In a preferred embodiment, Figure 10 is the 3rd untreated predeterminable area according to the preferred embodiment of the invention
Schematic diagram.As shown in Figure 10 it is assumed that the size of facial image is M × N, the size of the first predeterminable area is M ' × N', and M ' is permissible
TakeWithBetween part, and N' can take the part between 1 to N, then the size of the 3rd predeterminable area is X × Y,
X can takeWithBetween part, and Y can takeArriveBetween part.
Step S14:Left eye region and right eye region is determined respectively in the 3rd predeterminable area;
In a preferred embodiment, left eye region is the 3rd predeterminable area left-halfArriveBetween part,
And right eye region is the 3rd predeterminable area right half partArriveBetween part.
Step S15:Count gray value in left eye region and right eye region respectively and be more than the 4th predetermined threshold value
The number of pixel;
Step S16:Choose the 5th predetermined threshold value N from 0 to 255, then start to 255 to terminate just choosing each successively from N
Integer, calculates gray value in left eye region and right eye region respectively and is equal to each pixel of positive integer chosen
Number and the product of this positive integer, and calculate the summation of whole result of product, then using the summation calculating respectively divided by phase
The left eye region answered or the area of right eye region, enter to the summation of left eye region and right eye region respectively
Row normalized, wherein, N is 0 or positive integer and N≤255;
It is in the preferred embodiment assumed that N takes 245, statistics gray value is pixel number when 245 is n1Individual, gray value
It is n for pixel number when 2462Individual, gray value is pixel number when 247 is n3Individual ..., then will to carry out for the first time
The number of the pixel equal to 245 and 245 product, second by the number of the pixel carrying out equal to 246 and 246 take advantage of
Long-pending, for the third time by the product of the number of the pixel carrying out equal to 247 and 247, the rest may be inferred, up to calculating and 255 take advantage of
Long-pending, then each result of product is added and then obtains summation, that is, the result of calculation of each product addition summation is 245 × n1+
246×n2+247×n3+ ..., area that then also will correspondingly divided by left eye region or right eye region(Example
As:Left eye region area isM is the length of facial image, N is the width of facial image)It is normalized place
Reason.
Step S17:Whether the number being more than the pixel of the 4th predetermined threshold value according to the gray value counting is more than the 6th
Whether predetermined threshold value and the normalization result calculating are more than the 7th predetermined threshold value, judge whether the glasses in facial image are anti-
Light.
In a preferred embodiment, when the number that the gray value counting is more than the pixel of the 4th predetermined threshold value is more than the 6th
The predetermined threshold value and normalization result that calculates is more than the 7th predetermined threshold value, you can judge the glasses reflection in facial image.
With reference to the preferred implementation shown in Figure 11, the above-mentioned process that is preferable to carry out is further described.
Figure 11 is the judgement side of glasses and glasses reflection in face recognition process according to the preferred embodiment of the invention
The flow chart of method.As shown in figure 11, the method may comprise steps of:
Step S1102:Judge whether facial image wears sunglasses, specifically may comprise steps of:
Step S11021:Size according to facial image intercepts the middle ocular comprising glasses(Be equivalent to above-mentioned
One predeterminable area);
Step S11022:Count above-mentioned middle ocular gray value and be less than predetermined threshold value(For example:50, be equivalent to above-mentioned
First predetermined threshold value)Pixel number;
Step S11023:The number calculating the pixel that this area grayscale value is less than 50 accounts for whole region pixel number
Ratio, if greater than threshold value T(For example:T can be with value for middle ocular areaBe equivalent to the above-mentioned second default threshold
Value)Then it is assumed that wearing sunglasses, otherwise then it is assumed that not wearing sunglasses, step S1104 can be continued executing with;
Step S1104:In the middle of judging, whether ocular wears black surround glasses, specifically may comprise steps of:
Step S11041:Homomorphic filtering process, enhancing contrast ratio are carried out to the middle ocular getting;
Step S11042:Gaussian filter is utilized to the middle ocular after the homomorphic filtering of step S11041 is processed
Carry out smothing filtering, eliminate the interference of noise, subsequently traveled through using neighborhood minimum wave filter every in this middle ocular
Individual pixel, prominent target;
Step S11043:By the greyscale transformation of the middle ocular after the disposal of gentle filter to [0,1], using system
Meter rectangular histogram is met corresponding gray value T during proportion0, and adopt T0To the middle eye area after the disposal of gentle filter
Domain carries out binary conversion treatment;
Step S11044:Mid portion is intercepted on the middle ocular after binary conversion treatment(Be equivalent to above-mentioned
Second predeterminable area), wherein it is possible to include:Connect the connector being stuck on the bridge of the nose between left eye picture frame and right eye picture frame, pass through
Calculate proportion shared by the mid portion of above-mentioned intercepting for the connector region and predetermined threshold value(Be equivalent to the above-mentioned 3rd to preset
Threshold value)It is compared, and then judge whether wear black surround glasses;When connector region is in the mid portion of above-mentioned intercepting
When interior proportion is more than predetermined threshold value, you can the type judging glasses is black surround glasses;
Step S1106:In the middle of judging, whether ocular wears other common spectacles in addition to sunglasses and black surround glasses,
Specifically may comprise steps of:
Step S11061:Using Sobel operator filtering, binary conversion treatment is carried out to middle ocular, obtains binaryzation
Image img_bw;
Step S11062:Morphological operation is carried out to img_bw, the less part of area in img_bw image is deleted, goes
Fall part interference region;Again closed operation is carried out to middle ocular, the spaced point closing on is connected into connected region, smooth figure
The profile of picture;
Step S11063:In the middle of the binaryzation later to morphological operation in step S11062, ocular extracts connection point
Amount, determines size and the position of connected component;
Step S11064:Exclusive PCR connected region, further exclusive PCR region, each connection is judged by mark value
Region position in the picture, if this connected region contains more pixel near image boundary position, deleting should
Connected region, if connected region contains less pixel, deletes this connected region;
Step S11065:Calculate integral projection, calculate the middle ocular of binaryzation respectively in the horizontal direction
Deng the value of the integral projection at three, and horizontal integral projection is more than zero number, and finds out this according to the size of integral projection
Horizontal length in middle ocular connected region the longest;
Step S11066:Whether comprehensive descision wears common spectacles, meets one of following three kinds of situations and can be judged as wearing
Wear common spectacles:
If situation one horizontal directionOr horizontal directionPlace integral projection value be more than 2 and connected region
Number is more than or equal to 4;
Situation two, the gray value of image central region are that 1 number is more than or equal to 5, and horizontal integral projection is more than zero
Number is more than the half of picturedeep, and the number of connected region is more than or equal to 2;
Situation three, horizontal length connected region the longest is more than the half of picturewide.
Step S1108:Judge whether glasses are reflective, specifically may comprise steps of:
Step S11081:Retroreflective regions are intercepted on middle ocular(Be equivalent to above-mentioned 3rd predeterminable area), wherein,
The coverage of retroreflective regions includes:Human eye area, and retroreflective regions are less than middle ocular;
Step S11082:Left eye region and right eye region is determined respectively in retroreflective regions;
Step S11083:Count left eye region respectively, gray value is more than 250 in right eye region(Be equivalent to
Above-mentioned 4th predetermined threshold value)Pixel number, the as number of reflective spot;
Step S11084:Statistics left eye region, the rectangular histogram of right eye region respectively, for reflective picture,
In rectangular histogram, the number of pixels of high grade grey level or frequency are relatively large, using the rectangular histogram of high grade grey level be multiplied by gray level by terms of
Reflective degree;For example:Above-mentioned 5th predetermined threshold value takes 245, starts to 255 to terminate to choose each positive integer successively from 245
Carry out product calculation respectively, finally calculate the summation of each result of product again, then utilize above-mentioned summation divided by left eye location
Domain, the area of right eye region, are normalized;
Step S11085:In being can determine whether according to the result in above-mentioned steps S11083 and step S11084 and given threshold
Between ocular whether reflective, that is, when the gray value counting be more than 250 pixel number be more than predetermined threshold value(Be equivalent to
Above-mentioned 6th predetermined threshold value)And the normalization result calculating is more than predetermined threshold value(Be equivalent to above-mentioned 7th predetermined threshold value), that is,
Can determine whether that glasses are reflective under light irradiation;
Step S1110:Output glasses type and/or the whether reflective judged result of glasses.
In a preferred embodiment, above-mentioned judged result can include but is not limited to one below:Sunglasses, black surround glasses, general
Logical glasses, black surround glasses and reflective, common spectacles and reflective.
Figure 12 is the structured flowchart of the identification device of glasses in recognition of face according to embodiments of the present invention.As Figure 12 institute
Show, in this recognition of face, the identification device of glasses can include:Acquisition module 10, for obtaining facial image, and is getting
Facial image in intercept the first predeterminable area, wherein, the coverage of the first predeterminable area includes:Human eye area and glasses
The some or all of region covering;Judge module 20, for judging the type of glasses according to the first predeterminable area.
Using device as shown in figure 12, solve to lack in correlation technique and judging that glasses are in face recognition process
The no reflective problem to the technical scheme that the type of human eye wearing spectacles is differentiated before, and then can quickly and accurately sentence
Break and wearing spectacles and reflective facial image, the state of eyes in the human eye position fixing process of recognition of face is classified, right
Human eye in different situations is positioned using different methods, thus improve speed and the accuracy of human eye positioning.
Preferably, as shown in figure 13, judge module 20 can include:First statistic unit 200, pre- first for counting
If gray value is less than the number of the pixel of the first predetermined threshold value in region;First judging unit 202, for according to less than first
The ratio of number of the pixel in the number of the pixel of predetermined threshold value and the first predeterminable area and the ratio of the second predetermined threshold value
Relatively result, judges whether the type of glasses is sunglasses.
Preferably, as shown in figure 13, judge module 20 can also include:First processing units 204, for default to first
Region carries out homomorphic filtering process;Second processing unit 206, for entering to the first predeterminable area after homomorphic filtering is processed
Row the disposal of gentle filter;3rd processing unit 208, for carrying out neighborhood to the first predeterminable area after the disposal of gentle filter
Mini-value filtering is processed;Fourth processing unit 210, for entering to the first predeterminable area after neighborhood minimum Filtering Processing
Row fixed proportion binary conversion treatment;First interception unit 212, for presetting in first after fixed proportion binary conversion treatment
Second predeterminable area is intercepted on region, wherein, the second predeterminable area is symmetrical with regard to face axis, the second predeterminable area bag
Include:Connector between left eye picture frame and right eye picture frame;Second judging unit 214, for according to connector region
The comparative result of proportion and the 3rd predetermined threshold value in two predeterminable areas, judges whether the type of glasses is black surround glasses.
Preferably, as shown in figure 13, judge module 20 can also include:5th processing unit 216, for default to first
Region carries out Sobel operator filtering process;6th processing unit 218, for first after Sobel operator filtering is processed
Predeterminable area carries out binary conversion treatment;Extraction unit 220, for carrying out shape to the first predeterminable area after binary conversion treatment
State is processed, and extracts one or more connected regions;3rd judging unit 222, for carrying out to one or more connected regions
Integral projection calculates, and judges that whether the type of glasses is the other kinds of glasses in addition to sunglasses or black surround glasses.
Preferably, as shown in figure 13, judge module 20, are additionally operable to the type judging glasses and for black surround glasses or remove
After other kinds of glasses outside sunglasses or black surround glasses, judge whether the glasses in described facial image are reflective, sentence
Disconnected module 20 can also include:Second interception unit 224, for the 3rd predeterminable area is intercepted on the first predeterminable area, wherein,
The coverage of the 3rd predeterminable area includes:Human eye area, and the 3rd predeterminable area is less than the first predeterminable area;Determining unit
226, for determining left eye region and right eye region in the 3rd predeterminable area respectively;Second statistic unit 228, uses
In counting the number that gray value in left eye region and right eye region is more than the pixel of the 4th predetermined threshold value respectively;The
Three statistic units 230, in gray value such as:Choose the 5th predetermined threshold value N from 0 to 255, then start to 255 end from N,
Choose each positive integer successively, calculate gray value in left eye region and right eye region respectively and be equal to each just selection
The number of the pixel of integer and the product of this positive integer, and calculate the summation of whole result of product, then using calculating
Summation, divided by the area of left eye region and right eye region, is normalized to summation, and wherein, N is 0 or just
Integer and N≤255;4th judging unit 232, is more than the pixel of the 4th predetermined threshold value for the gray value that basis counts
Number, whether more than the 6th predetermined threshold value and whether the normalization result that calculates is more than the 7th predetermined threshold value, judges face figure
Whether the glasses in picture are reflective under light irradiation.
As can be seen from the above description, above embodiments enable following technique effect(It should be noted that these
Effect is the effect that some preferred embodiments can reach):Technical scheme provided by the present invention can quickly and accurately judge
Go out to wear glasses and reflective facial image, the state of eyes in human eye positioning is classified, for the human eye of different situations
Positioned using different methods, thus improve speed and the accuracy of human eye positioning.
Obviously, those skilled in the art should be understood that each module of the above-mentioned present invention or each step can be with general
Computing device realizing, they can concentrate on single computing device, or be distributed in multiple computing devices and formed
Network on, alternatively, they can be realized with the executable program code of computing device, it is thus possible to they are stored
To be executed by computing device in the storage device, and in some cases, can be with different from shown in order execution herein
The step going out or describing, or they are fabricated to respectively each integrated circuit modules, or by the multiple modules in them or
Step is fabricated to single integrated circuit module to realize.So, the present invention is not restricted to any specific hardware and software combination.
The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for the skill of this area
For art personnel, the present invention can have various modifications and variations.All within the spirit and principles in the present invention, made any repair
Change, equivalent, improvement etc., should be included within the scope of the present invention.
Claims (8)
1. in a kind of recognition of face glasses discrimination method it is characterised in that include:
Obtain facial image, and intercept the first predeterminable area in the described facial image getting, wherein, described first presets
The coverage in region includes:Human eye area and some or all of region of glasses covering;
Judge the type of described glasses according to described first predeterminable area;
Wherein, judge that the type of described glasses includes according to described first predeterminable area:Rope is carried out to described first predeterminable area
Bel's Sobel operator filtering is processed;The first predeterminable area after described Sobel operator filtering is processed is carried out at binaryzation
Reason;The first predeterminable area after described binary conversion treatment is carried out with Morphological scale-space, extracts one or more connected regions;
One or more of connected regions are integrated with projection calculate, judge that whether the type of described glasses is except sunglasses or black
Other kinds of glasses outside frame glasses.
2. method according to claim 1 it is characterised in that judge the class of described glasses according to described first predeterminable area
Type includes:
Statistics gray value in described first predeterminable area is less than the number of the pixel of the first predetermined threshold value;
The number of the pixel in the number according to the pixel less than described first predetermined threshold value and described first predeterminable area
Ratio and the second predetermined threshold value comparative result, judge whether the type of described glasses is sunglasses.
3. method according to claim 1 it is characterised in that judge the class of described glasses according to described first predeterminable area
Type includes:
Homomorphic filtering process is carried out to described first predeterminable area;
The disposal of gentle filter is carried out to the first predeterminable area after described homomorphic filtering is processed;
Neighborhood minimum Filtering Processing is carried out to the first predeterminable area after described the disposal of gentle filter;
Ratio binary conversion treatment is fixed to the first predeterminable area after described neighborhood minimum Filtering Processing;
The first predeterminable area after described fixed proportion binary conversion treatment intercepts the second predeterminable area, wherein, described
Second predeterminable area is symmetrical with regard to face axis, and described second predeterminable area includes:Left eye picture frame and right eye picture frame it
Between connector;
Comparison knot according to described connector region proportion and the 3rd predetermined threshold value in described second predeterminable area
Really, whether the type judging described glasses is black surround glasses.
4. the method according to claim 3 or 1 is it is characterised in that be described black surround in the type judging described glasses
After glasses or the other kinds of glasses in addition to sunglasses or described black surround glasses, judge the eye in described facial image
The whether reflective inclusion of mirror:
Described first predeterminable area intercepts the 3rd predeterminable area, wherein, the coverage of described 3rd predeterminable area includes:
Human eye area, and described 3rd predeterminable area is less than described first predeterminable area;
Determine left eye region and right eye region respectively in described 3rd predeterminable area;
Count the pixel that gray value in described left eye region and described right eye region is more than the 4th predetermined threshold value respectively
The number of point;
Choose the 5th predetermined threshold value N from 0 to 255, then start to 255 to terminate to choose each positive integer successively from N, calculate respectively
In described left eye region and described right eye region, gray value is equal to the number of each pixel of positive integer chosen
With the product of this positive integer, and calculate the summation of whole result of product, then using the described summation calculating divided by a described left side
Eye region and the area of described right eye region, are normalized to described summation, and wherein, N is 0 or just whole
Number and N≤255;
Whether the number being more than the pixel of the 4th predetermined threshold value according to the described gray value counting is more than the 6th predetermined threshold value
And whether the described normalization result calculating is more than the 7th predetermined threshold value, judge whether the glasses in described facial image are anti-
Light.
5. in a kind of recognition of face glasses identification device it is characterised in that include:
Acquisition module, for obtaining facial image, and intercepts the first predeterminable area in the described facial image getting, its
In, the coverage of described first predeterminable area includes:Human eye area and some or all of region of glasses covering;
Judge module, for judging the type of described glasses according to described first predeterminable area;
Wherein, described judge module includes:5th processing unit, for carrying out Sobel Sobel to described first predeterminable area
Operator filtering is processed;6th processing unit, for carrying out to the first predeterminable area after described Sobel operator filtering is processed
Binary conversion treatment;Extraction unit, for carrying out Morphological scale-space to the first predeterminable area after described binary conversion treatment, carries
Take one or more connected regions;3rd judging unit, for being integrated projection meter to one or more of connected regions
Calculate, judge that whether the type of described glasses is the other kinds of glasses in addition to sunglasses or black surround glasses.
6. device according to claim 5 is it is characterised in that described judge module includes:
First statistic unit, for statistics, in described first predeterminable area, gray value is less than the pixel of the first predetermined threshold value
Number;
First judging unit, for the number according to the pixel less than described first predetermined threshold value and described first predeterminable area
The ratio of the number of interior pixel and the comparative result of the second predetermined threshold value, judge whether the type of described glasses is sunglasses.
7. device according to claim 5 is it is characterised in that described judge module includes:
First processing units, for carrying out homomorphic filtering process to described first predeterminable area;
Second processing unit, for carrying out the disposal of gentle filter to the first predeterminable area after described homomorphic filtering is processed;
3rd processing unit, for carrying out neighborhood minimum filtering to the first predeterminable area after described the disposal of gentle filter
Process;
Fourth processing unit, for being fixed ratio to the first predeterminable area after described neighborhood minimum Filtering Processing
Binary conversion treatment;
First interception unit, pre- for intercepting second on the first predeterminable area after described fixed proportion binary conversion treatment
If region, wherein, described second predeterminable area is symmetrical with regard to face axis, and described second predeterminable area includes:Left eye
Connector between picture frame and right eye picture frame;
Second judging unit, for according to described connector region in described second predeterminable area proportion and the 3rd
The comparative result of predetermined threshold value, judges whether the type of described glasses is black surround glasses.
8. the device according to claim 7 or 5, it is characterised in that described judge module, is additionally operable to judging described eye
After the type of mirror is described black surround glasses or the other kinds of glasses in addition to sunglasses or described black surround glasses, judge
Whether the glasses in described facial image are reflective, and described judge module also includes:
Second interception unit, for intercepting the 3rd predeterminable area, wherein, described 3rd preset areas on described first predeterminable area
The coverage in domain includes:Human eye area, and described 3rd predeterminable area is less than described first predeterminable area;
Determining unit, for determining left eye region and right eye region respectively in described 3rd predeterminable area;
Second statistic unit, is more than for counting in described left eye region and described right eye region gray value respectively
The number of the pixel of four predetermined threshold value;
3rd statistic unit, for choosing the 5th predetermined threshold value N from 0 to 255, then starts to 255 to terminate to choose often successively from N
Individual positive integer, calculates gray value in described left eye region and described right eye region respectively and is equal to the just whole of each selection
The number of pixel of number and the product of this positive integer, and calculate the summation of whole result of product, then using the institute calculating
State the area divided by described left eye region and described right eye region for the summation, described summation be normalized,
Wherein, N is 0 or positive integer and N≤255;
4th judging unit, for be more than according to the described gray value that counts the 4th predetermined threshold value pixel number whether
More than the 6th predetermined threshold value and whether the described normalization result that calculates is more than the 7th predetermined threshold value, judge described face figure
Whether the glasses in picture are reflective.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310027751.8A CN103093210B (en) | 2013-01-24 | 2013-01-24 | Method and device for glasses identification in face identification |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310027751.8A CN103093210B (en) | 2013-01-24 | 2013-01-24 | Method and device for glasses identification in face identification |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103093210A CN103093210A (en) | 2013-05-08 |
CN103093210B true CN103093210B (en) | 2017-02-08 |
Family
ID=48205758
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310027751.8A Active CN103093210B (en) | 2013-01-24 | 2013-01-24 | Method and device for glasses identification in face identification |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103093210B (en) |
Families Citing this family (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103465825A (en) * | 2013-08-15 | 2013-12-25 | 苏州佳世达光电有限公司 | Vehicle-mounted system and control method thereof |
CN105095829B (en) * | 2014-04-29 | 2019-02-19 | 华为技术有限公司 | A kind of face identification method and system |
CN106155280A (en) * | 2015-03-30 | 2016-11-23 | 北京智谷睿拓技术服务有限公司 | Exchange method, interactive device and subscriber equipment |
CN105335695A (en) * | 2015-08-19 | 2016-02-17 | 华南理工大学 | Glasses detection based eye positioning method |
CN105184253B (en) * | 2015-09-01 | 2020-04-24 | 北京旷视科技有限公司 | Face recognition method and face recognition system |
CN106503611B (en) * | 2016-09-09 | 2019-11-22 | 西安理工大学 | Facial image eyeglass detection method based on marginal information projective iteration mirror holder crossbeam |
CN107808120B (en) * | 2017-09-30 | 2018-08-31 | 平安科技(深圳)有限公司 | Glasses localization method, device and storage medium |
CN108564540B (en) * | 2018-03-05 | 2020-07-17 | Oppo广东移动通信有限公司 | Image processing method and device for removing lens reflection in image and terminal equipment |
CN108492421A (en) * | 2018-03-29 | 2018-09-04 | 成都惠网远航科技有限公司 | Low-power consumption face identification method |
CN108600555A (en) * | 2018-06-15 | 2018-09-28 | 努比亚技术有限公司 | A kind of screen color method of adjustment, mobile terminal and computer readable storage medium |
CN109036332A (en) * | 2018-09-14 | 2018-12-18 | 奇酷互联网络科技(深圳)有限公司 | Intelligent terminal and its adjusting method of screen intensity, the device with store function |
CN109657652A (en) * | 2019-01-16 | 2019-04-19 | 平安科技(深圳)有限公司 | A kind of face identification method and device |
CN109902561A (en) * | 2019-01-16 | 2019-06-18 | 平安科技(深圳)有限公司 | A kind of face identification method and device, robot applied to robot |
CN111179880A (en) * | 2019-12-26 | 2020-05-19 | 恒大新能源汽车科技(广东)有限公司 | Brightness adjusting method and device of display screen, electronic equipment and system |
CN111488843A (en) * | 2020-04-16 | 2020-08-04 | 贵州安防工程技术研究中心有限公司 | Face sunglasses distinguishing method based on step-by-step inhibition of missing report and false report rate |
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 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101499127A (en) * | 2008-02-03 | 2009-08-05 | 上海银晨智能识别科技有限公司 | Method for preventing trouble in human face recognition caused by interference |
CN102034079A (en) * | 2009-09-24 | 2011-04-27 | 汉王科技股份有限公司 | Method and system for identifying faces shaded by eyeglasses |
CN102262727A (en) * | 2011-06-24 | 2011-11-30 | 常州锐驰电子科技有限公司 | Method for monitoring face image quality at client acquisition terminal in real time |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102324166B (en) * | 2011-09-19 | 2013-06-12 | 深圳市汉华安道科技有限责任公司 | Fatigue driving detection method and device |
-
2013
- 2013-01-24 CN CN201310027751.8A patent/CN103093210B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101499127A (en) * | 2008-02-03 | 2009-08-05 | 上海银晨智能识别科技有限公司 | Method for preventing trouble in human face recognition caused by interference |
CN102034079A (en) * | 2009-09-24 | 2011-04-27 | 汉王科技股份有限公司 | Method and system for identifying faces shaded by eyeglasses |
CN102262727A (en) * | 2011-06-24 | 2011-11-30 | 常州锐驰电子科技有限公司 | Method for monitoring face image quality at client acquisition terminal in real time |
Non-Patent Citations (1)
Title |
---|
Eyeglasses Removal From Facial Image Based On Phase Congruency;Xiaodong Jia,et al.;《2010 3rd International Congress on Image and Signal Processing(CISP2010)》;20101231;第1859-1862页 * |
Also Published As
Publication number | Publication date |
---|---|
CN103093210A (en) | 2013-05-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103093210B (en) | Method and device for glasses identification in face identification | |
CN108038456B (en) | Anti-deception method in face recognition system | |
CN103902977B (en) | Face identification method and device based on Gabor binary patterns | |
EP1991948B1 (en) | An iris recognition system having image quality metrics | |
RU2431190C2 (en) | Facial prominence recognition method and device | |
CN105373779B (en) | A kind of vehicle safety belt intelligent detecting method and intelligent checking system | |
CN103902962B (en) | One kind is blocked or the adaptive face identification method of light source and device | |
CN109635656A (en) | Vehicle attribute recognition methods, device, equipment and medium neural network based | |
CN109657652A (en) | A kind of face identification method and device | |
CN105893920A (en) | Human face vivo detection method and device | |
CN101615292B (en) | Accurate positioning method for human eye on the basis of gray gradation information | |
CN103020579A (en) | Face recognition method and system, and removing method and device for glasses frame in face image | |
CN103942539B (en) | A kind of oval accurate high efficiency extraction of head part and masking method for detecting human face | |
CN109858439A (en) | A kind of biopsy method and device based on face | |
CN103473564B (en) | A kind of obverse face detection method based on sensitizing range | |
CN103383723A (en) | Method and system for spoof detection for biometric authentication | |
CN103902958A (en) | Method for face recognition | |
CN109858471A (en) | Biopsy method, device and computer equipment based on picture quality | |
CN108428214A (en) | A kind of image processing method and device | |
CN110728185B (en) | Detection method for judging existence of handheld mobile phone conversation behavior of driver | |
CN105426843A (en) | Single-lens palm vein and palmprint image acquisition apparatus and image enhancement and segmentation method | |
Martinikorena et al. | Fast and robust ellipse detection algorithm for head-mounted eye tracking systems | |
CN115039150A (en) | Determination method, determination device, and determination program | |
CN106920266B (en) | The Background Generation Method and device of identifying code | |
EP3961495A1 (en) | System and method for finding an area of an eye from a facial image |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |