CN102375970A - Identity authentication method based on face and authentication apparatus thereof - Google Patents
Identity authentication method based on face and authentication apparatus thereof Download PDFInfo
- Publication number
- CN102375970A CN102375970A CN2010102542016A CN201010254201A CN102375970A CN 102375970 A CN102375970 A CN 102375970A CN 2010102542016 A CN2010102542016 A CN 2010102542016A CN 201010254201 A CN201010254201 A CN 201010254201A CN 102375970 A CN102375970 A CN 102375970A
- Authority
- CN
- China
- Prior art keywords
- face
- people
- facial image
- attitude
- posture
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000001514 detection method Methods 0.000 claims abstract description 7
- 230000001815 facial effect Effects 0.000 claims description 78
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 230000015572 biosynthetic process Effects 0.000 claims description 2
- 230000003247 decreasing effect Effects 0.000 abstract 1
- 210000003128 head Anatomy 0.000 description 23
- 230000008859 change Effects 0.000 description 9
- 206010034719 Personality change Diseases 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000012549 training Methods 0.000 description 4
- 210000005252 bulbus oculi Anatomy 0.000 description 3
- 238000004141 dimensional analysis Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 210000000887 face Anatomy 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 230000005764 inhibitory process Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000000452 restraining effect Effects 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
Images
Landscapes
- Collating Specific Patterns (AREA)
Abstract
The invention provides an identity authentication method based on a face and an authentication apparatus thereof. The method comprises the following steps: carrying out face detection on each frame of image acquired by a camera device to obtain a face image, and tracking the face image to obtain a tracking result; carrying out an attitude analysis, wherein if a face attitude in obtained attitude information is not jumped relative to a face attitude corresponding to a prior frame of image, an attitude analysis result is not jumped; carrying out face comparison; carrying out statistics analysis on a plurality of frames of tracking results, a plurality of frames of attitude analysis results and a plurality of frames of face image authentication results, wherein if a statistics analysis result is in accordance with a preset condition, identity authentication is passed. According to the method and the apparatus thereof in the invention, a two-dimensional flat picture is utilized to distinguish a photograph from a human, in the prior art utilizing three-dimensional information and carrying out photograph identification cause that a system identification speed is substantially decreased and an application requirement can not satisfied, and the above technical problem is solved in the invention.
Description
Technical field
The present invention relates to the identity identifying technology of facial image, particularly relate to a kind of identity identifying method and authenticate device based on people's face.
Background technology
Along with the development of face recognition technology, the application relevant with recognition of face increases gradually, and the face authentication system has obtained increasing application as the application system of face recognition technology at aspects such as automatic gate inhibition, the logins of people's face.
The face authentication system utilizes camera collection personnel's to be certified facial image, compares with the facial image of corresponding identity in the information bank, if comparison is passed through; Think that then the facial image of corresponding identity has identical identity in person to be certified and the storehouse; Authentication is passed through, otherwise authentication is not passed through.For authentication is passed through, the photo that the jactitator possibly hold personnel in the information bank carries out authentication, if Verification System can not be distinguished photo and true man, will make jactitator's authentication of holding photo pass through.Therefore, need in system, increase the function of distinguishing photo and true man, differentiation true man and photo are one of face authentication system problems that need solve.
Photo and true man's difference mainly contains: photo is two-dimentional, and true man are three-dimensional.Utilize this difference,, comprise methods such as the binocular image synthesizes through the reconstruct of three-dimensional face; Can distinguish photo and true man, but the three-dimensional data amount is very big, computing velocity is slow; The binocular camera shooting head such as also need calibrate at operation, and therefore, the resource that this method not only need take is more; Also cause system identification speed to descend greatly, and be difficult to satisfy real-time application demand.
Summary of the invention
The purpose of this invention is to provide a kind of identity identifying method and authenticate device based on people's face; Can utilize 2 dimensional plane image areas to tell photo and true man, the solution prior art is utilized three-dimensional information and is carried out photo and discern the technical matters that the system identification speed that is caused descends, can't satisfy application requirements greatly.
To achieve these goals, on the one hand, a kind of identity identifying method based on people's face is provided, has comprised:
Every two field picture of camera head collection is carried out people's face detect and obtain facial image, said facial image is followed the tracks of the acquisition tracking results, if said tracking results is for tracing into people's face, corresponding people's face belongs to same people in then said people's face and the previous frame image;
Said facial image is carried out posture analysis, obtain attitude information, if saltus step does not take place corresponding human face posture in the relative previous frame image of the human face posture in the said attitude information, then the posture analysis result is not saltus step;
Image in the storehouse of said facial image and appointment is carried out the comparison of people's face; If comparison result be similar or similarity greater than assign thresholds; Think that this facial image is the people in the database of appointment, the authentication result of then said present frame facial image is for passing through, otherwise for not passing through;
The said tracking results of multiframe, the said posture analysis result of multiframe and the said facial image authentication result of multiframe are carried out statistical study, if statistic analysis result is consistent with prerequisite, then authentication is passed through.
Preferably, in the above-mentioned method, said prerequisite is:
Said tracking results be trace into people's face and said posture analysis result for the continuous frame number of not saltus step greater than first thresholding; And,
Said facial image authentication result is that to trace into people's face and said posture analysis result be that the ratio of the not continuous frame number of saltus step is greater than second thresholding for the frame number that passes through and said tracking results; And,
Said tracking results be trace into people's face and said posture analysis result for the variable quantity of the human face posture in the continuous multiple frames of not saltus step greater than the 3rd thresholding.
Preferably, in the above-mentioned method, said prerequisite is:
Keep the posture analysis result not the human face posture in the 1st frame to the n frame of the said facial image of saltus step meet the first attitude scope; And,
Keep the posture analysis result not the human face posture in n+k frame to the n+k+m frame of the said facial image of saltus step meet the second attitude scope; Wherein, n, m are the integer greater than 1, and k is the integer more than or equal to 1.
Preferably, in the above-mentioned method, before said every two field picture to the camera head collection carries out step that people's face detect to follow the tracks of, also comprise: through the conversion head pose as requested of the tested personnel before voice or the picture cues camera lens.
Preferably, in the above-mentioned method, also comprise: if said tracking results is not for tracing into people's face, said present frame face tracking breaks off, and then removes said statistic analysis result, restarts statistics; The generation saltus step as a result of perhaps said posture analysis, said present frame Attitude Tracking is broken off, and then removes said statistic analysis result, restarts statistics.
Preferably; In the above-mentioned method; Said facial image is carried out in the step of posture analysis; Specifically comprise: in said facial image, measure formation first distance between first key point and second key point, measure between the 3rd key point and the 4th key point and form second distance, with the ratio of said first distance and said second distance as attitude parameter; If the variable quantity of the attitude parameter of the relative previous frame image of the attitude parameter of said facial image changes threshold value less than predetermined attitude, the attitude saltus step does not take place in then said facial image.
Preferably, in the above-mentioned method, said first key point is that left nare, said second key point are the left side corners of the mouth, and said the 3rd key point is a right naris, and said the 4th key point is the right side corners of the mouth.
To achieve these goals, the embodiment of the invention also provides a kind of identification authentication system based on people's face, comprising:
People's face detects tracking module; Be used for: the every two field picture to the camera head collection carries out people's face detection acquisition facial image; Said facial image is followed the tracks of; Obtain tracking results, if said tracking results is for tracing into people's face, corresponding people's face belongs to same people in then said people's face and the previous frame image;
The posture analysis module is used for: said facial image is carried out posture analysis, obtain attitude information, if saltus step does not take place corresponding human face posture in the relative previous frame image of the human face posture in the said attitude information, then the posture analysis result is not saltus step;
The face authentication module; Be used for: the storehouse image of said facial image and appointment is carried out the comparison of people's face; If comparison result be similar or similarity greater than assign thresholds; Think that this facial image is the people in the database of appointment, the authentication result of then said present frame facial image is for passing through, otherwise for not passing through;
The comprehensive module of multiframe authentication result is used for: the said tracking results of multiframe, the said posture analysis result of multiframe and the said facial image authentication result of multiframe are carried out statistical study, if statistic analysis result is consistent with prerequisite, then authentication is passed through.
Preferably, in the above-mentioned device, said prerequisite is: said tracking results be trace into people's face and said posture analysis result for the continuous frame number of not saltus step greater than first thresholding; And said facial image authentication result is that to trace into people's face and said posture analysis result be that the ratio of the not continuous frame number of saltus step is greater than second thresholding for the frame number that passes through and said tracking results; And, said tracking results be trace into people's face and said posture analysis result for the variable quantity of the human face posture in the continuous multiple frames of not saltus step greater than the 3rd thresholding; Perhaps,
Said prerequisite is: keep the posture analysis result not the human face posture in the 1st frame to the n frame of the said facial image of saltus step meet the first attitude scope; And, keep the posture analysis result not the human face posture in n+k frame to the n+k+m frame of the said facial image of saltus step meet the second attitude scope; Wherein, n, m are the integer greater than 1, and k is the integer more than or equal to 1.
Preferably, in the above-mentioned device, also comprise:
Prompting and authentication result output module are used for: through the conversion head pose as requested of the tested personnel before voice or the picture cues camera lens; Export said tested personnel's authentication result through voice or image.
Preferably, in the above-mentioned device, comprise also that the replacement module is used for:
If said tracking results is not for tracing into people's face, said present frame face tracking breaks off, and then removes the statistic analysis result of the comprehensive module of said multiframe authentication result, makes the comprehensive module of said multiframe authentication result restart statistics; Saltus step for taking place in perhaps said posture analysis result, and said present frame Attitude Tracking is broken off, and then removes the statistic analysis result of the comprehensive module of said multiframe authentication result, makes the comprehensive module of said multiframe authentication result restart statistics.
There is following technique effect at least in the present invention:
1) the present invention combines through face tracking, Attitude Tracking, posture analysis and face authentication, can improve authentication performance, avoid the jactitator with photo through authentication.
2) embodiment of the invention is not carried out three dimensional analysis; Only from the angle analysis shooting angle of the two dimension difference of facial image simultaneously not, also can judge the number of people and change with respect to the angle of filming apparatus, and; Two-dimensional image data is abundant; Image Acquisition is easy, and simultaneously, data volume is much smaller than the three-dimensional face data.
Description of drawings
The flow chart of steps of the method that Fig. 1 provides for the embodiment of the invention;
Fig. 2 carries out the processed steps process flow diagram for what the embodiment of the invention provided to present frame;
The structural drawing of the device that Fig. 3 provides for the embodiment of the invention;
The head anglec of rotation synoptic diagram that Fig. 4 provides for the embodiment of the invention;
The authentication sample collection synoptic diagram that Fig. 5 provides for the embodiment of the invention.
Embodiment
For the purpose, technical scheme and the advantage that make the embodiment of the invention is clearer, will combine accompanying drawing that specific embodiment is described in detail below.
The flow chart of steps of the method that Fig. 1 provides for the embodiment of the invention, as shown in Figure 1, utilize plane picture to carry out the method for face authentication, comprising:
The embodiment of the invention is not carried out three dimensional analysis; Only from the angle analysis shooting angle of the two dimension difference of facial image simultaneously not, also can judge the number of people and change with respect to the angle of filming apparatus, and; Two-dimensional image data is abundant; Image Acquisition is easy, and simultaneously, data volume is much smaller than the three-dimensional face data.
Wherein, said prerequisite can for: said tracking results be trace into people's face and said posture analysis result for the continuous frame number of not saltus step greater than first thresholding; And said facial image authentication result is that to trace into people's face and said posture analysis result be that the ratio of the not continuous frame number of saltus step is greater than second thresholding for the frame number that passes through and said tracking results; And, said tracking results be trace into people's face and said posture analysis result for the variable quantity of the human face posture in the continuous multiple frames of not saltus step greater than the 3rd thresholding.
Said prerequisite can also for: keep the posture analysis result not the human face posture in the 1st frame to the n frame of the said facial image of saltus step meet the first attitude scope; And, keep the posture analysis result not the human face posture in n+k frame to the n+k+m frame of the said facial image of saltus step meet the second attitude scope; Wherein, n, m are the integer greater than 1, and k is the integer more than or equal to 1.
Wherein, if said tracking results is not for tracing into people's face, said present frame face tracking breaks off, and then removes said statistic analysis result, restarts statistics; Saltus step for taking place in perhaps said posture analysis result, and said present frame Attitude Tracking is broken off, and then removes said statistic analysis result, restarts statistics.
Said face tracking image is carried out in the step 102 of posture analysis; Specifically comprise: in said facial image, measure and form first distance between first key point and second key point; Measure between the 3rd key point and the 4th key point and form second distance, with the ratio of said first distance and said second distance as attitude parameter; If the variable quantity of the attitude parameter of the relative previous frame image of the attitude parameter of said facial image changes threshold value less than predetermined attitude, the attitude saltus step does not take place in then said facial image.For example, said first key point is that left nare, said second key point are the left side corners of the mouth, and said the 3rd key point is a right naris, and said the 4th key point is the right side corners of the mouth.
Wherein, before step 101, can also comprise: through the conversion head pose as requested of the tested personnel before voice or the picture cues camera lens.The conversion head pose can for about shake the head or nod up and down.
It is thus clear that; The embodiment of the invention provides a kind of face authentication method that detects tracking and posture analysis based on people's face; The requirement user makes the action of head left rotation and right rotation in verification process, thereby variation has taken place the head pose of the feasible image that collects, because photo can't carry out the head rotation; Therefore, the method can be distinguished single photo and true man.Simultaneously, utilize face tracking, can confirm to take place that attitude changes is same individual face, if utilize two different photos of attitude to carry out authentication, people's face can't be followed the tracks of during the conversion photo, attitude also can't follow the tracks of.Thereby distinguish true man and photo.
Fig. 2 carries out the processed steps process flow diagram for what the embodiment of the invention provided to present frame, and is as shown in Figure 2, comprising:
Whether step 203 is judged that people's face detect to be followed the tracks of, and is execution in step 204 then, otherwise execution in step 206;
Step 208 is carried out the attitude of current frame image and is classified and face authentication;
The embodiment of the invention also provides a kind of device of face authentication, the structural drawing of the device that Fig. 3 provides for the embodiment of the invention, and it comprises:
People's face detects tracking module 301, and be used for: the every two field picture to the camera head collection carries out the detection and tracking of people's face, obtains facial image, if tracking results is for tracing into people's face, corresponding people's face belongs to same people in then said people's face and the previous frame image;
Posture analysis module 302 is used for: said facial image is carried out posture analysis, obtain attitude information, if saltus step does not take place corresponding human face posture in the relative previous frame image of the human face posture in the said attitude information, then the posture analysis result is not saltus step;
Face authentication module 303; Be used for: the storehouse image of said facial image and appointment is carried out the comparison of people's face; If comparison result be similar or similarity greater than assign thresholds; Think that this facial image is the people in the database of appointment, the authentication result of then said present frame facial image is for passing through, otherwise for not passing through;
The comprehensive module 304 of multiframe authentication result is used for: the said tracking results of multiframe, the said posture analysis result of multiframe and the said facial image authentication result of multiframe are carried out statistical study, if statistic analysis result is consistent with prerequisite, then authentication is passed through.
Wherein, can also comprise: prompting and authentication result output module are used for: through the conversion head pose as requested of the tested personnel before voice or the picture cues camera lens; Export said tested personnel's authentication result through voice or image.
Can also comprise; The replacement module is used for: if said tracking results is not for tracing into people's face, said present frame face tracking breaks off; Then remove the statistic analysis result of the comprehensive module of said multiframe authentication result, make the comprehensive module of said multiframe authentication result restart statistics; Saltus step for taking place in perhaps said posture analysis result, and said present frame Attitude Tracking is broken off, and then removes the statistic analysis result of the comprehensive module of said multiframe authentication result, makes the comprehensive module of said multiframe authentication result restart statistics.
Below, each module is elaborated.
People's face detects tracking module 301.
The detection and tracking of people's face belong to the technology of present comparative maturity.People's face detects method (the Adaboost algorithms: a kind of iterative algorithm that adopt based on the Adaboost algorithm more; Its core concept is to the different Weak Classifier of same training set training; Gather these Weak Classifiers then; Constitute a stronger final sorter), through the sorter of a large amount of people's faces and non-face image training of human face.Face tracking has based on the method for MeanShift algorithm (Mean Shift algorithm: the step that generally is meant an iteration; Promptly calculate the skew average of current point earlier; Move this to its skew average,, continue to move then as new starting point; Up to the end that meets some requirements .), based on the method for statistical model etc., face tracking can be used for following the tracks of in the video same people's face.If the performance of track algorithm is fine, if so in the present frame detected people's face next frame do not disappear, then can be followed the tracks of.If people's face of several frames in front and back can not think that the people's face in the video no longer is same individual face on following the tracks of.Thereby, can confirm through face tracking whether detected people's face is same individual face in people's face and the former frame in the present frame.Gone up if present frame is followed the tracks of, explained that people's face and the people's face in the former frame in the present frame is same individual face, passed through, then directly provided the signal that authentication is passed through, no longer carried out follow-up processing if former frame this person face is authenticated.If former frame this person does not have authentication to pass through, possibly be this person data not enough, can not satisfy the requirement that authentication is passed through, then this person's face is inputed to posture analysis and face authentication module, continue follow-up processing.On if present frame is not followed the tracks of; Explain that then detected in the present frame is new people's face; The data message that then the last people's appearance in the comprehensive module of multiframe authentication result is closed is removed, and then this person's face is inputed to posture analysis and face authentication module, continues follow-up processing.
Posture analysis module 302.
The target of posture analysis is to analyze when the angle of forefathers' appearance for camera, in tracing process, has carried out the head rotation if work as forefathers' face, thinks that then working as forefathers' face is not photo.Because the attitude of single photo is confirmed, if multiple pictures even attitude is different between the photo, also can be considered to new people's face appearance owing to not following the tracks of when photo switches, thereby new person's face still has only a kind of attitude.Here, multiple pictures refers to more than two or two.
The head anglec of rotation synoptic diagram that Fig. 4 provides for the embodiment of the invention, as can be seen from the figure, head can be rotated around three axles.Around the rotation of Z axle is rotation in the plane, does not change the relation of people's face imaging surface and camera, therefore, can not distinguish photo and true man.Around X axle rotation, because nose etc. with respect to the angle of camera variation have taken place, the nose change of lower direction above that is not easy to detect; And around Y axle when rotation; Nose can form shade in both sides and block, and people's face left and right sides also can form shade and block simultaneously, and the variation of people's face is bigger; Therefore, select the variation of this direction to distinguish true man and photo.
Posture analysis can obtain through the relation of training attitude and imaging; It is continuous that attitude changes; That is to say that around axis, head can be rotated with continuous angle, thus when making the camera imaging and the direction of front face have the continually varying angle.In this case, can attitude simply be divided into three kinds in front, left side, right side.
When attitude changed, bigger variation can take place in the face structure of people's face, made the detection and tracking of people's face relatively more difficult, also reduced the performance of face authentication.Can follow the tracks of for underwriter's face, need guarantee that the attitude variation can not be excessive.Here can adopt the restraining line, both sides to limit personnel's to be certified head rotation amplitude.
For applied environment of the present invention; Need not confirm the accurate angle of people's appearance for camera; Only need know the degree that the attitude of the image of different frame changes; Therefore, the present invention proposes the method that a kind of simple and effective sign attitude changes, and utilizes the let others have a look at attitude of face of the mutual alignment relation table of a plurality of key points of people's face.People's key point on the face has eyeball, canthus, nose, nostril, the corners of the mouth, and wherein, eyeball is rotary; The position has bigger variation; And the existence of glasses can influence the location at eyeball and canthus, and nose is located relatively difficulty in non-direct picture; Nostril, the corners of the mouth are necessarily can be detected, but possibly have only a corners of the mouth or a nostril.Here; Utilize the relative position relation of the nostril and the corners of the mouth and they and other key points to represent the attitude of people's face; Concrete grammar is: the distance of the definition left nare and the left side corners of the mouth is a, and the distance of the right naris and the right side corners of the mouth is b, and the ratio of a and b can change during the head rotation; And during the photo left rotation and right rotation, the ratio of a and b does not change.Attitude and about the definition See Figure.
Can carry out the simple classification of attitude with the ratio of a and b.
That is: a/b＜1 left side
A/b=1 faces
A/b＞1 right side
Through key point, also can confirm other similarly rules.Perhaps carry out through the attitude sorter.At first, attitude is divided into three types in front, left side, right side,, therefore, does not select pitching here because the variation photo of pitching also can realize.Then, select a large amount of people's face samples for every kind of attitude, need consider environment of applications when sample is selected, like the requirement of face tracking, the sample of selection all must be able to trace into.Then sample is carried out feature extraction; As extract Gabor (breakaway poing) characteristic, ASM (active shape model), LBP (Local Binary Patterns; Local binary pattern) etc., and carry out characteristic and select, can adopt the method for Adaboost to carry out characteristic and select; Train the attitude sorter then, commonly used is SVM (SVMs) sorter.Utilize the attitude sorter, can carry out the attitude classification given facial image.
Because the rotation of people's head is continuous, therefore, the variation of the ratio of a/b should be satisfied certain rule, and promptly the difference of the ratio of front and back frame should be not excessive, and certainly, this is under the situation that guarantees certain frame per second, to obtain.If the ratio difference between two frames of front and back surpasses certain threshold value, think then the attitude saltus step taken place that on promptly attitude was not followed the tracks of, two frames did not belong to same individual face before and after thinking.
Face authentication module 303.
Face authentication is that the identity corresponding sample that facial image and the user with input selects is compared, if same type then authentication pass through, otherwise authentication is not passed through.The user select the method for identity can be swipe the card, method such as click (also can not select identity, all data are compared respectively in people's face to be identified like this and the storehouse, if having one through authentication authentication through).The flow process of face authentication does, carries out feature extraction earlier, then proper vector inputed to sorter, confirms according to the output result of sorter whether authentication is passed through.Characteristic commonly used has PCA dimensionality reduction characteristic, Gabor characteristic, LBP characteristic, histogram feature etc.; If desired; Can also after feature extraction, carry out selecting of characteristic, sorter commonly used have boosting sorter, svm classifier device, Bayes classifier, type in type between sorter etc.Through face authentication, can know that whether current facial image is through authentication.
The comprehensive module 304 of multiframe authentication result.
The comprehensive module of multiframe authentication result is one of main contents of the present invention, also is the principal element that influences the authentication effect.Here, two kinds of possible strategy are proposed:
Strategy one satisfies the authentication of following formula and passes through.Wherein, th1, th2, th3 are preset threshold value, can be according to having experience and actual application environment to be provided with.As, the performance of face authentication algorithm is fine, and the tracking frame number that then needs is less, and the authentication frame number is higher with the ratio of following the tracks of frame number, and attitude changes requirement greatly.
Follow the tracks of frame number＞th1;
Authentication frame number/tracking frame number＞th2;
Attitude variation >=th3.
Wherein, following the tracks of that frame number refers to is the number that face tracking and Attitude Tracking are all followed the tracks of
Strategy two adopts to have the sample collection method of prompting.The authentication sample collection synoptic diagram that Fig. 5 provides for the embodiment of the invention, as shown in Figure 5, if 1-n frame sample satisfies that attitude 1 requires and authentication satisfies to m frame sample through, n+1 that attitude 2 requires and authentication is passed through, then the authentication of this person's face is passed through.Require the attitude number to have two kinds at least, as adopting front face and right side or left side.Reminding method can be methods such as voice, screen picture.Perhaps carry out sample collection according to the order of left, center, right.
Strategy one two respectively has relative merits with strategy, can select according to authentication method etc.Strategy one does not need prompting, and person to be certified can carry out authentication naturally, changes attitude naturally; And the facial image that strategy one identification module can a kind of attitude of authentication only need be supported front face like identification module; Also can support multiple attitude, but need multiframe with the assurance authentication performance, and; Not prompting does not change if the user knows to carry out attitude, then can't be through authentication.Strategy two needs prompting, needs sorter to support multiple attitude, because prompting is arranged, can make things convenient for the user to cooperate, and have the face authentication of two or more attitudes to pass through, more misclassification rates that reduce more.
After authentication is passed through, the authentication output result.
The authentication result output module, the output of authentication result can be adopted voice, auditory tone cues, picture cues etc.
By on can know that the embodiment of the invention has following advantage:
1) the present invention combines through face tracking, Attitude Tracking, posture analysis and face authentication, can improve authentication performance, avoid the jactitator with photo through authentication.
2) embodiment of the invention is not carried out three dimensional analysis; Only from the angle analysis shooting angle of the two dimension difference of facial image simultaneously not, also can judge the number of people and change with respect to the angle of filming apparatus, and; Two-dimensional image data is abundant; Image Acquisition is easy, and simultaneously, data volume is much smaller than the three-dimensional face data.
The above only is a preferred implementation of the present invention; Should be pointed out that for those skilled in the art, under the prerequisite that does not break away from the principle of the invention; Can also make some improvement and retouching, these improvement and retouching also should be regarded as protection scope of the present invention.
Claims (11)
1. the identity identifying method based on people's face is characterized in that, comprising:
Every two field picture of camera head collection is carried out people's face detect and obtain facial image, said facial image is followed the tracks of the acquisition tracking results, if said tracking results is for tracing into people's face, corresponding people's face belongs to same people in then said people's face and the previous frame image;
Said facial image is carried out posture analysis, obtain attitude information, if saltus step does not take place corresponding human face posture in the relative previous frame image of the human face posture in the said attitude information, then the posture analysis result is not saltus step;
Image in the storehouse of said facial image and appointment is carried out the comparison of people's face; If comparison result be similar or similarity greater than assign thresholds; Think that this facial image is the people in the database of appointment, the authentication result of then said present frame facial image is for passing through, otherwise for not passing through;
The said tracking results of multiframe, the said posture analysis result of multiframe and the said facial image authentication result of multiframe are carried out statistical study, if statistic analysis result is consistent with prerequisite, then authentication is passed through.
2. identity identifying method according to claim 1 is characterized in that, said prerequisite is:
Said tracking results be trace into people's face and said posture analysis result for the continuous frame number of not saltus step greater than first thresholding; And,
Said facial image authentication result is that to trace into people's face and said posture analysis result be that the ratio of the not continuous frame number of saltus step is greater than second thresholding for the frame number that passes through and said tracking results; And,
Said tracking results be trace into people's face and said posture analysis result for the variable quantity of the human face posture in the continuous multiple frames of not saltus step greater than the 3rd thresholding.
3. identity identifying method according to claim 1 is characterized in that, said prerequisite is:
Keep the posture analysis result not the human face posture in the 1st frame to the n frame of the said facial image of saltus step meet the first attitude scope; And,
Keep the posture analysis result not the human face posture in n+k frame to the n+k+m frame of the said facial image of saltus step meet the second attitude scope; Wherein, n, m are the integer greater than 1, and k is the integer more than or equal to 1.
4. according to claim 1,2 or 3 described identity identifying methods; It is characterized in that; Before said every two field picture to the camera head collection carries out step that people's face detect to follow the tracks of, also comprise: through the conversion head pose as requested of the tested personnel before voice or the picture cues camera lens.
5. identity identifying method according to claim 1 is characterized in that, also comprises: if said tracking results is not for tracing into people's face, said present frame face tracking breaks off, and then removes said statistic analysis result, restarts statistics; The generation saltus step as a result of perhaps said posture analysis, said present frame Attitude Tracking is broken off, and then removes said statistic analysis result, restarts statistics.
6. identity identifying method according to claim 1; It is characterized in that; Said facial image is carried out in the step of posture analysis; Specifically comprise: in said facial image, measure formation first distance between first key point and second key point, measure between the 3rd key point and the 4th key point and form second distance, with the ratio of said first distance and said second distance as attitude parameter; If the variable quantity of the attitude parameter of the relative previous frame image of the attitude parameter of said facial image changes threshold value less than predetermined attitude, the attitude saltus step does not take place in then said facial image.
7. identity identifying method according to claim 6 is characterized in that, said first key point is that left nare, said second key point are the left side corners of the mouth, and said the 3rd key point is a right naris, and said the 4th key point is the right side corners of the mouth.
8. the identification authentication system based on people's face is characterized in that, comprising:
People's face detects tracking module; Be used for: the every two field picture to the camera head collection carries out people's face detection acquisition facial image; Said facial image is followed the tracks of; Obtain tracking results, if said tracking results is for tracing into people's face, corresponding people's face belongs to same people in then said people's face and the previous frame image;
The posture analysis module is used for: said facial image is carried out posture analysis, obtain attitude information, if saltus step does not take place corresponding human face posture in the relative previous frame image of the human face posture in the said attitude information, then the posture analysis result is not saltus step;
The face authentication module; Be used for: the storehouse image of said facial image and appointment is carried out the comparison of people's face; If comparison result be similar or similarity greater than assign thresholds; Think that this facial image is the people in the database of appointment, the authentication result of then said present frame facial image is for passing through, otherwise for not passing through;
The comprehensive module of multiframe authentication result is used for: the said tracking results of multiframe, the said posture analysis result of multiframe and the said facial image authentication result of multiframe are carried out statistical study, if statistic analysis result is consistent with prerequisite, then authentication is passed through.
9. identification authentication system according to claim 8 is characterized in that,
Said prerequisite is: said tracking results be trace into people's face and said posture analysis result for the continuous frame number of not saltus step greater than first thresholding; And said facial image authentication result is that to trace into people's face and said posture analysis result be that the ratio of the not continuous frame number of saltus step is greater than second thresholding for the frame number that passes through and said tracking results; And, said tracking results be trace into people's face and said posture analysis result for the variable quantity of the human face posture in the continuous multiple frames of not saltus step greater than the 3rd thresholding; Perhaps,
Said prerequisite is: keep the posture analysis result not the human face posture in the 1st frame to the n frame of the said facial image of saltus step meet the first attitude scope; And, keep the posture analysis result not the human face posture in n+k frame to the n+k+m frame of the said facial image of saltus step meet the second attitude scope; Wherein, n, m are the integer greater than 1, and k is the integer more than or equal to 1.
10. identification authentication system according to claim 8 is characterized in that, also comprises:
Prompting and authentication result output module are used for: through the conversion head pose as requested of the tested personnel before voice or the picture cues camera lens; Export said tested personnel's authentication result through voice or image.
11. identification authentication system according to claim 8 is characterized in that, also comprises, the replacement module is used for:
If said tracking results is not for tracing into people's face, said present frame face tracking breaks off, and then removes the statistic analysis result of the comprehensive module of said multiframe authentication result, makes the comprehensive module of said multiframe authentication result restart statistics; Saltus step for taking place in perhaps said posture analysis result, and said present frame Attitude Tracking is broken off, and then removes the statistic analysis result of the comprehensive module of said multiframe authentication result, makes the comprehensive module of said multiframe authentication result restart statistics.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010254201.6A CN102375970B (en) | 2010-08-13 | 2010-08-13 | A kind of identity identifying method based on face and authenticate device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010254201.6A CN102375970B (en) | 2010-08-13 | 2010-08-13 | A kind of identity identifying method based on face and authenticate device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102375970A true CN102375970A (en) | 2012-03-14 |
| CN102375970B CN102375970B (en) | 2016-03-30 |
Family
ID=45794557
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010254201.6A Active CN102375970B (en) | 2010-08-13 | 2010-08-13 | A kind of identity identifying method based on face and authenticate device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102375970B (en) |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102779274A (en) * | 2012-07-19 | 2012-11-14 | 冠捷显示科技（厦门）有限公司 | Intelligent television face recognition method based on binocular camera |
| CN103020607A (en) * | 2012-12-27 | 2013-04-03 | Tcl集团股份有限公司 | Face recognition method and face recognition device |
| CN103500330A (en) * | 2013-10-23 | 2014-01-08 | 中科唯实科技（北京）有限公司 | Semi-supervised human detection method based on multi-sensor and multi-feature fusion |
| CN103605969A (en) * | 2013-11-28 | 2014-02-26 | Tcl集团股份有限公司 | Method and device for face inputting |
| CN104348619A (en) * | 2013-07-31 | 2015-02-11 | 联想(北京)有限公司 | Identity verifying method and terminal device |
| CN104778390A (en) * | 2014-01-10 | 2015-07-15 | 由田新技股份有限公司 | Identity authentication system and method thereof |
| CN104794458A (en) * | 2015-05-07 | 2015-07-22 | 北京丰华联合科技有限公司 | Fuzzy video person identifying method |
| CN105407098A (en) * | 2015-11-26 | 2016-03-16 | 小米科技有限责任公司 | Identity verification method and device |
| CN105868677A (en) * | 2015-01-19 | 2016-08-17 | 阿里巴巴集团控股有限公司 | Live human face detection method and device |
| CN105989338A (en) * | 2015-02-13 | 2016-10-05 | 多媒体影像解决方案有限公司 | Face recognition method and system thereof |
| CN106203242A (en) * | 2015-05-07 | 2016-12-07 | 阿里巴巴集团控股有限公司 | A kind of similar image recognition methods and equipment |
| CN106682591A (en) * | 2016-12-08 | 2017-05-17 | 广州视源电子科技股份有限公司 | Face recognition method and device |
| CN106778198A (en) * | 2016-11-23 | 2017-05-31 | 北京小米移动软件有限公司 | Perform the safety certifying method and device of operation |
| WO2017101267A1 (en) * | 2015-12-18 | 2017-06-22 | 腾讯科技（深圳）有限公司 | Method for identifying living face, terminal, server, and storage medium |
| CN108021846A (en) * | 2016-11-01 | 2018-05-11 | 杭州海康威视数字技术股份有限公司 | A kind of face identification method and device |
| CN108921145A (en) * | 2018-08-31 | 2018-11-30 | 深圳市研本品牌设计有限公司 | Based on hot spot character news method for tracing and system |
| CN109034130A (en) * | 2018-08-31 | 2018-12-18 | 深圳市研本品牌设计有限公司 | A kind of unmanned plane and storage medium for news tracking |
| WO2019080580A1 (en) * | 2017-10-26 | 2019-05-02 | 深圳奥比中光科技有限公司 | 3d face identity authentication method and apparatus |
| WO2019080579A1 (en) * | 2017-10-26 | 2019-05-02 | 深圳奥比中光科技有限公司 | 3d face identity authentication method and apparatus |
| WO2019080578A1 (en) * | 2017-10-26 | 2019-05-02 | 深圳奥比中光科技有限公司 | 3d face identity authentication method and apparatus |
| CN110008673A (en) * | 2019-03-06 | 2019-07-12 | 阿里巴巴集团控股有限公司 | A kind of identification authentication method and apparatus based on recognition of face |
| CN111091028A (en) * | 2018-10-23 | 2020-05-01 | 北京嘀嘀无限科技发展有限公司 | Method and device for recognizing shaking motion and storage medium |
| CN111783677A (en) * | 2020-07-03 | 2020-10-16 | 北京字节跳动网络技术有限公司 | Face recognition method, face recognition device, server and computer readable medium |
| US10997722B2 (en) | 2018-04-25 | 2021-05-04 | Beijing Didi Infinity Technology And Development Co., Ltd. | Systems and methods for identifying a body motion |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1694045A (en) * | 2005-06-02 | 2005-11-09 | 北京中星微电子有限公司 | Non-contact type visual control operation system and method |
| CN101216887A (en) * | 2008-01-04 | 2008-07-09 | 浙江大学 | An automatic computer authentication method for photographic faces and living faces |
| WO2009078957A1 (en) * | 2007-12-14 | 2009-06-25 | Flashfoto, Inc. | Systems and methods for rule-based segmentation for objects with full or partial frontal view in color images |
| CN101710383A (en) * | 2009-10-26 | 2010-05-19 | 北京中星微电子有限公司 | Method and device for identity authentication |
| 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 |
-
2010
- 2010-08-13 CN CN201010254201.6A patent/CN102375970B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1694045A (en) * | 2005-06-02 | 2005-11-09 | 北京中星微电子有限公司 | Non-contact type visual control operation system and method |
| WO2009078957A1 (en) * | 2007-12-14 | 2009-06-25 | Flashfoto, Inc. | Systems and methods for rule-based segmentation for objects with full or partial frontal view in color images |
| CN101216887A (en) * | 2008-01-04 | 2008-07-09 | 浙江大学 | An automatic computer authentication method for photographic faces and living faces |
| CN101710383A (en) * | 2009-10-26 | 2010-05-19 | 北京中星微电子有限公司 | Method and device for identity authentication |
| 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 |
Cited By (36)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102779274B (en) * | 2012-07-19 | 2015-02-25 | 冠捷显示科技（厦门）有限公司 | Intelligent television face recognition method based on binocular camera |
| CN102779274A (en) * | 2012-07-19 | 2012-11-14 | 冠捷显示科技（厦门）有限公司 | Intelligent television face recognition method based on binocular camera |
| CN103020607B (en) * | 2012-12-27 | 2017-05-03 | Tcl集团股份有限公司 | Face recognition method and face recognition device |
| CN103020607A (en) * | 2012-12-27 | 2013-04-03 | Tcl集团股份有限公司 | Face recognition method and face recognition device |
| CN104348619A (en) * | 2013-07-31 | 2015-02-11 | 联想(北京)有限公司 | Identity verifying method and terminal device |
| CN104348619B (en) * | 2013-07-31 | 2018-12-14 | 联想(北京)有限公司 | Verify the method and terminal device of identity |
| CN103500330A (en) * | 2013-10-23 | 2014-01-08 | 中科唯实科技（北京）有限公司 | Semi-supervised human detection method based on multi-sensor and multi-feature fusion |
| CN103605969A (en) * | 2013-11-28 | 2014-02-26 | Tcl集团股份有限公司 | Method and device for face inputting |
| CN103605969B (en) * | 2013-11-28 | 2018-10-09 | Tcl集团股份有限公司 | A kind of method and device of face typing |
| CN104778390A (en) * | 2014-01-10 | 2015-07-15 | 由田新技股份有限公司 | Identity authentication system and method thereof |
| CN105868677A (en) * | 2015-01-19 | 2016-08-17 | 阿里巴巴集团控股有限公司 | Live human face detection method and device |
| CN105989338A (en) * | 2015-02-13 | 2016-10-05 | 多媒体影像解决方案有限公司 | Face recognition method and system thereof |
| CN104794458A (en) * | 2015-05-07 | 2015-07-22 | 北京丰华联合科技有限公司 | Fuzzy video person identifying method |
| CN106203242A (en) * | 2015-05-07 | 2016-12-07 | 阿里巴巴集团控股有限公司 | A kind of similar image recognition methods and equipment |
| CN106203242B (en) * | 2015-05-07 | 2019-12-24 | 阿里巴巴集团控股有限公司 | Similar image identification method and equipment |
| CN105407098A (en) * | 2015-11-26 | 2016-03-16 | 小米科技有限责任公司 | Identity verification method and device |
| WO2017101267A1 (en) * | 2015-12-18 | 2017-06-22 | 腾讯科技（深圳）有限公司 | Method for identifying living face, terminal, server, and storage medium |
| CN106897658A (en) * | 2015-12-18 | 2017-06-27 | 腾讯科技（深圳）有限公司 | The discrimination method and device of face live body |
| US10438077B2 (en) | 2015-12-18 | 2019-10-08 | Tencent Technology (Shenzhen) Company Limited | Face liveness detection method, terminal, server and storage medium |
| CN106897658B (en) * | 2015-12-18 | 2021-12-14 | 腾讯科技（深圳）有限公司 | Method and device for identifying human face living body |
| CN108021846A (en) * | 2016-11-01 | 2018-05-11 | 杭州海康威视数字技术股份有限公司 | A kind of face identification method and device |
| CN106778198A (en) * | 2016-11-23 | 2017-05-31 | 北京小米移动软件有限公司 | Perform the safety certifying method and device of operation |
| CN106682591B (en) * | 2016-12-08 | 2020-04-07 | 广州视源电子科技股份有限公司 | Face recognition method and device |
| CN106682591A (en) * | 2016-12-08 | 2017-05-17 | 广州视源电子科技股份有限公司 | Face recognition method and device |
| WO2019080579A1 (en) * | 2017-10-26 | 2019-05-02 | 深圳奥比中光科技有限公司 | 3d face identity authentication method and apparatus |
| WO2019080578A1 (en) * | 2017-10-26 | 2019-05-02 | 深圳奥比中光科技有限公司 | 3d face identity authentication method and apparatus |
| WO2019080580A1 (en) * | 2017-10-26 | 2019-05-02 | 深圳奥比中光科技有限公司 | 3d face identity authentication method and apparatus |
| US11238270B2 (en) | 2017-10-26 | 2022-02-01 | Orbbec Inc. | 3D face identity authentication method and apparatus |
| US10997722B2 (en) | 2018-04-25 | 2021-05-04 | Beijing Didi Infinity Technology And Development Co., Ltd. | Systems and methods for identifying a body motion |
| CN109034130A (en) * | 2018-08-31 | 2018-12-18 | 深圳市研本品牌设计有限公司 | A kind of unmanned plane and storage medium for news tracking |
| CN108921145A (en) * | 2018-08-31 | 2018-11-30 | 深圳市研本品牌设计有限公司 | Based on hot spot character news method for tracing and system |
| CN111091028A (en) * | 2018-10-23 | 2020-05-01 | 北京嘀嘀无限科技发展有限公司 | Method and device for recognizing shaking motion and storage medium |
| CN110008673A (en) * | 2019-03-06 | 2019-07-12 | 阿里巴巴集团控股有限公司 | A kind of identification authentication method and apparatus based on recognition of face |
| CN110008673B (en) * | 2019-03-06 | 2022-02-18 | 创新先进技术有限公司 | Identity authentication method and device based on face recognition |
| CN111783677A (en) * | 2020-07-03 | 2020-10-16 | 北京字节跳动网络技术有限公司 | Face recognition method, face recognition device, server and computer readable medium |
| CN111783677B (en) * | 2020-07-03 | 2023-12-01 | 北京字节跳动网络技术有限公司 | Face recognition method, device, server and computer readable medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102375970B (en) | 2016-03-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102375970A (en) | Identity authentication method based on face and authentication apparatus thereof | |
| Shao et al. | Deep convolutional dynamic texture learning with adaptive channel-discriminability for 3D mask face anti-spoofing | |
| KR101877981B1 (en) | System for recognizing disguised face using gabor feature and svm classifier and method thereof | |
| Mahbub et al. | Partial face detection for continuous authentication | |
| CN105740842B (en) | Unsupervised face identification method based on fast density clustering algorithm | |
| CN107346422B (en) | Living body face recognition method based on blink detection | |
| CN102385703B (en) | A kind of identity identifying method based on face and system | |
| Tayab Khan et al. | Smart real-time video surveillance platform for drowsiness detection based on eyelid closure | |
| CN105447432A (en) | Face anti-fake method based on local motion pattern | |
| CN104008364A (en) | Face recognition method | |
| CN102880864A (en) | Method for snap-shooting human face from streaming media file | |
| HN et al. | Human Facial Expression Recognition from static images using shape and appearance feature | |
| JP2013228847A (en) | Facial expression analyzing device and facial expression analyzing program | |
| Lee et al. | Distinction Between Real Faces and Photos by Analysis of Face Data. | |
| Tofighi et al. | A robust face recognition system in image and video | |
| CN104751144A (en) | Frontal face quick evaluation method for video surveillance | |
| Juang et al. | Vision-based human body posture recognition using support vector machines | |
| Bakheet | A fuzzy framework for real-time gesture spotting and recognition | |
| Gul et al. | A machine learning approach to detect occluded faces in unconstrained crowd scene | |
| Paul et al. | Automatic adaptive facial feature extraction using CDF analysis | |
| Singh et al. | Comparison analysis of different face detecting techniques | |
| Ying et al. | An automatic system for multi-view face detection and pose estimation | |
| Angulu et al. | Landmark localization approach for facial computing | |
| Wang et al. | Finger-fist detection in first-person view based on monocular vision using Haar-like features | |
| Reddy et al. | Driver drowsiness monitoring based on eye map and mouth contour |
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 | ||
| C41 | Transfer of patent application or patent right or utility model | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20160517 Address after: 519031 Guangdong city of Zhuhai province Hengqin Baohua Road No. 6, room 105 -478 Patentee after: GUANGDONG ZHONGXING ELECTRONICS CO., LTD. Address before: 100083, Haidian District, Xueyuan Road, Beijing No. 35, Nanjing Ning building, 15 Floor Patentee before: Beijing Vimicro Corporation |