CN106156688A - A kind of dynamic human face recognition methods and system - Google Patents
A kind of dynamic human face recognition methods and system Download PDFInfo
- Publication number
- CN106156688A CN106156688A CN201510102708.2A CN201510102708A CN106156688A CN 106156688 A CN106156688 A CN 106156688A CN 201510102708 A CN201510102708 A CN 201510102708A CN 106156688 A CN106156688 A CN 106156688A
- Authority
- CN
- China
- Prior art keywords
- face
- personnel
- optimum
- dynamic human
- face recognition
- 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.)
- Pending
Links
Landscapes
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a kind of dynamic human face recognition methods, if described method include extracting face characteristic data with set up key population data base, carry out monitor video whether the moving object of moving body track, detecting and tracking exists face and be ranked up, extract face characteristic data according to optimum facial image, the face characteristic data of acquisition and the face characteristic data in key population data base being compared finds have the face of high similarity with personnel in storehouse, sends the steps such as information.The invention also discloses a kind of dynamic human face recognition system, described system includes DBM, motion tracking module, image collection module, characteristic extracting module, feature comparing module and reminding module, further, including optimum face judge module and camera driver module.The dynamic human face recognition methods of present invention offer and system, the face characteristic that can extract in monitor video is compared, and is sent information when the result that discovery similarity is higher.
Description
Technical field
The present invention relates to technical field of face recognition, particularly relate to a kind of dynamic human face recognition methods
And system.
Background technology
Due to anti-terrorism, Homeland Security, the needs of social safety, every country is all to peace in the world
Step up its investment in anti-field.And a key problem of identification security protection just.But at mesh
In front identification procedure, main employing is all mode of manually interrogating and examining, and such mode is only capable of
Little scope crowd uses, once enters the crowd's body under big flow crowd or specific environment
Part the most substantially exposes its shortcoming if identifying: waste of manpower resource, accuracy of identification is limited, pipe
Reason difficulty, it is difficult to ensure to identify rapidly identity information accurately and effectively, and make respective handling and arrange
Execute.
In recent years, along with developing rapidly of computer technology, the automatic identification technology of biological characteristic
It is widely studied and develops, such as fingerprint recognition, palm shape identification etc..Recognition of face is based on face
The identity of picture differentiates, it is intended to makes computer have the function discriminated one's identification by face picture, is one
Plant the intelligence system relying on the high-tech such as image understanding, pattern recognition, computer vision, with
Other human body biological characteristics identification technology is compared has feature direct, friendly, convenient, is
Naturally direct means, it is easy to accepted by user.Face recognition technology application prospect is extensive, can
For bank, the monitoring system of customs and automatic gatekeeper system etc..Particularly at contactless environment
In the case of letting alone detected person, the superiority of face recognition technology is considerably beyond existing
The detection methods such as fingerprint.Compare to other biological identification technology, the advantage of face recognition technology
Clearly.First, as identifying feature, face has stable, reliable, safe and convenient
Etc. feature.In the ordinary course of things, the facial characteristics of people is the most reliable and the most stable, " carrying "
Convenient.Image surface is also the primary feature for distinguishing people.Just because of this, in public security, safety
The investigation of department, in security, access and exit control, portrait is widely used.Secondly,
The collection of facial image is very convenient, is based especially on the image acquisition mode of normal video.
But, currently for the video monitoring system of public security safety-security area, its main application is also
The video image being dependent on collecting carries out analysis afterwards, and early warning and prevention and control are limited in one's ability in advance,
Have a greatly reduced quality to the efficiency of clear up a criminal case.Therefore, Public Security Organss at different levels and state security department
Hitting all kinds of crimes, during maintaining social stability and national security, advanced person need introduced
And effective technological means, the General Promotion investigation ability to all kinds of cases.
Summary of the invention
In view of the present situation of current identification, the present invention provide a kind of dynamic human face recognition methods and
System, can extract the optimum face characteristic in monitor video and carry out real-time comparison, and find phase
Point out in time during like the comparison result that degree is higher.
For reaching above-mentioned purpose, embodiments of the invention adopt the following technical scheme that
A kind of dynamic human face recognition methods, described dynamic human face recognition methods comprises the following steps:
Extract face characteristic data to set up key population data base;
Real-time dynamic monitoring video is carried out moving body track;
Whether the moving object of detecting and tracking exists face and is ranked up;
Face characteristic data inputting facial feature database is extracted according to optimum facial image;
The face characteristic data of acquisition are carried out with the face characteristic data in key population data base
Comparison;
If discovery has the personnel of high similarity with personnel in key population data base, send prompting
Information.
According to one aspect of the present invention, whether the moving object of described detecting and tracking exists people
Face is also ranked up concretely: whether there is face, and root in the moving object of detecting and tracking
According to movement locus and the contour feature of face, obtain human-face detector by great amount of samples training,
Optimum facial image sequence is carried out according to face eyes opening and closing situation, direction of visual lines and facial orientation.
According to one aspect of the present invention, described dynamic human face recognition methods is further comprising the steps of:
If finding there are the personnel of high similarity repeatedly in same monitoring with personnel in key population data base
Region is identified to, then can assert that these personnel hover in this region, and send information.
According to one aspect of the present invention, described dynamic human face recognition methods is further comprising the steps of:
If find several with key population data base in personnel there are the personnel of high similarity in same monitoring
Region is identified to, then send information.
According to one aspect of the present invention, described according to optimum facial image and to extract face special
Levy data to include:
Reducing broad image, attitude is corrected;
Carry out eigenvalue location;
Behind location, picture is normalized;
Eigenvalue is calculated according to the picture after processing.
According to one aspect of the present invention, described extract face characteristic according to optimum facial image
Data inputting facial feature database includes:
Follow the tracks of optimum facial image;
Use Inline Function and fixed-point calculation that the facial image intercepted is carried out positioning feature point and spy
Value indicative is extracted;
The face characteristic value typing facial feature database that will extract.
According to one aspect of the present invention, find and personnel's tool in key population data base if described
The personnel having high similarity then send information concretely: if finding and key population data
In storehouse, personnel have the personnel of high similarity, then indicate in monitored picture, show this people
The photo of member and associated personal information, send the information of specific rank.
A kind of dynamic human face recognition system, described dynamic human face recognition system includes:
DBM, is used for extracting face characteristic data to set up key population data base;
Motion tracking module, for carrying out moving body track to real-time dynamic monitoring video;
Whether image collection module, exist face in the moving object of detecting and tracking and carry out
Sequence;
Characteristic extracting module, for extracting face characteristic data inputting according to optimum facial image
Facial feature database;
Feature comparing module, in the face characteristic data that will obtain and key population data base
Face characteristic data compare;
Reminding module, for finding have high similarity with personnel in key population data base
Information is sent during personnel.
According to one aspect of the present invention, described image collection module includes: optimum face judges
Module, for the movement locus according to face and contour feature, is trained by great amount of samples
To human-face detector, carry out according to face eyes opening and closing situation, direction of visual lines and facial orientation
Excellent facial image sorts.
According to one aspect of the present invention, described extract face characteristic according to optimum facial image
Data inputting facial feature database includes:
Follow the tracks of optimum facial image;
Use Inline Function and fixed-point calculation that the facial image intercepted is carried out positioning feature point and spy
Value indicative is extracted;
The face characteristic value typing facial feature database that will extract.
The advantage that the present invention implements: dynamic human face recognition methods of the present invention includes following step
Rapid: to extract face characteristic data to set up key population data base;To real-time dynamic monitoring video
Carry out moving body track;Whether the moving object of detecting and tracking exists face and is ranked up;
Face characteristic data inputting facial feature database is extracted according to optimum facial image;To obtain
Face characteristic data compare with the face characteristic data in key population data base;If sending out
The personnel now with personnel in key population data base with high similarity then send information, can
, in the case of coordinating without personnel, to automatically extract the optimum face in monitor video, warehouse-in
Preserve and carry out real-time comparison with backstage key population data base, finding the comparison that similarity is higher
Result is pointed out in time, meets the monitoring field management and control efficient to all kinds of personnel of public security protection video
Application demand.
Accompanying drawing explanation
For the technical scheme being illustrated more clearly that in the embodiment of the present invention, below will be to embodiment
The accompanying drawing used required in is briefly described, it should be apparent that, the accompanying drawing in describing below
It is only some embodiments of the present invention, for those of ordinary skill in the art, is not paying
On the premise of going out creative work, it is also possible to obtain other accompanying drawing according to these accompanying drawings.
Fig. 1 is a kind of dynamic human face recognition methods embodiment one schematic diagram of the present invention;
Fig. 2 is the structural representation of a kind of dynamic human face recognition system of the present invention;
Fig. 3 is a kind of dynamic human face recognition methods embodiment two schematic diagram of the present invention.
Detailed description of the invention
Below in conjunction with the accompanying drawing in the embodiment of the present invention, to the technical side in the embodiment of the present invention
Case is clearly and completely described, it is clear that described embodiment is only the present invention one
Divide embodiment rather than whole embodiments.Based on the embodiment in the present invention, this area is general
The every other embodiment that logical technical staff is obtained under not making creative work premise,
Broadly fall into the scope of protection of the invention.
A kind of dynamic human face recognition methods embodiment one
As it is shown in figure 1, a kind of dynamic human face recognition methods, described dynamic human face recognition methods bag
Include following steps:
Step S1: extract face characteristic data to set up key population data base;
Described step S1 extracts face characteristic data to set up being embodied as of key population data base
Mode can be: by obtaining related personnel's identity information with national public security system, household register system,
And personal information in key population data base, obtain all personnel by existing photograph image etc.
Face characteristic data, set up demographic data storehouse, set up for personnel in key population data base
Key population data base.
In actual applications, in described key population data base, personnel can be high-risk group, such as
The fugitive personnel in the whole nation, terrorist and previous conviction personnel etc., can be set to black by these high-risk group
List.
In actual applications, in described key population data base, personnel can be the crowd that passes through especially,
The management personnel in such as certain building, the current personnel of indult, People's Armed Police, public security etc., can
This kind of crowd is set to white list.
Step S2: real-time dynamic monitoring video is carried out moving body track;
Described step S2 carries out the specific embodiment party of moving body track to real-time dynamic monitoring video
Formula can be: by observing real-time dynamic monitoring video, the moving object to occurring in video is carried out
Follow the tracks of.
In actual applications, described real-time dynamic monitoring video is by DSP video camera or network
IP high-definition camera is monitored production, can take the photograph at described DSP video camera or network IP high definition
Camera performs step S4.
Step S3: whether there is face in the moving object of detecting and tracking and be ranked up;
Whether the moving object of described step S3 detecting and tracking exists face the tool being ranked up
Body embodiment can be: whether there is face in the moving object of detecting and tracking, and according to face
Movement locus and contour feature, by great amount of samples training obtain human-face detector, according to
Face eyes opening and closing situation, direction of visual lines and facial orientation carry out optimum facial image sequence.
In actual applications, optimum face can be with China second-generation identity card certificate photo as standard, just
Face is amimia, uniform light, without fuzzy, and the facial image of the eyeball that keeps one's eyes open.
In actual applications, the identification of optimum face can include following manner:
Facial feature localization----judge expression;
Attitude Algorithm----judge facial orientation;
Illumination algorithm----judge that luminosity is the most uniform, it is high light or backlight;
Fuzziness judges---whether image obscures.In actual applications, in the motion of detecting and tracking
When whether object exists face, move by following the tracks of target object, detect whether it is personnel,
Then proceed to detect the face whether these personnel occur in video, the most just face is tracked
Judge optimum face.
In actual applications, in the moving object of detecting and tracking, whether there is face and arrange
During sequence, can automatically lock picture occur everyone, the most anti-for the people hovered in picture
Multiple identification, at utmost reduces interference.
Step S4: extract face characteristic data inputting face characteristic number according to optimum facial image
According to storehouse;
Described step S4 extracts face characteristic data inputting face characteristic according to optimum facial image
Data base specifically can comprise the following steps that
Follow the tracks of optimum facial image;
Use Inline Function and fixed-point calculation that the facial image intercepted is carried out positioning feature point and spy
Value indicative is extracted;
The face characteristic value typing facial feature database that will extract.
In actual applications, according to optimum facial image and extract face characteristic data and comprise the steps that
Reducing broad image, attitude is corrected;
Carry out eigenvalue location, position including face, facial feature localization etc.;
Behind location, picture is normalized;
Calculate eigenvalue according to the picture after processing, specifically can pass through geometry location, use small echo
The modes such as conversion calculate eigenvalue.
In actual applications, eigenvalue comprises the steps that geometrical relationship;Color;Specific characteristic, than
Such as the nevus of face, birthmark etc..
In actual applications, the detailed description of the invention following the tracks of optimum facial image can be: to optimum
Facial image carries out Image semantic classification;Pretreated image is carried out Face datection;To detection
The face gone out carries out algorithm keeps track.
In actual applications, described employing Inline Function and the fixed-point calculation facial image to intercepting
Carrying out positioning feature point and characteristics extraction is high at DSP video camera or network IP when being embodied as
Clear video camera performs, in wherein Inline Function is the hardware that described video camera chips is corresponding
Connection function.
In actual applications, the described face characteristic value typing facial feature database that will extract
Before execution, described video camera needs the face characteristic value extracted and other face information data
It is transferred to server with typing facial feature database.
In actual applications, described employing Inline Function and the fixed-point calculation facial image to intercepting
The process that implements carrying out positioning feature point and characteristics extraction can be as follows:
A kind of facial modeling combined based on geometric projection and template matching is used to calculate
Method;First sciagraphy coarse positioning eye position is used;Then in this result, use PCA template
Matching method is accurately positioned;Finally according to the position location of eyes, use the 2 of sciagraphy location nose
Individual angle point and nose.
The key step of feature extraction is as follows:
Measurement Relation extraction eyebrow according to face and eyes window;
To eyebrow and eyes window inner projection coarse positioning eye position;The eyebrow and the eyes window that obtain are
Rectangle frame, if rectangle frame left upper apex coordinate and the coordinate of bottom right vertex, closes according to projection function
System, in calculation block any point in the horizontal direction with the average gray value in vertical direction, eyebrow
Hair and eyeball are black compared with other location comparisons, and gray value is in the horizontal direction at eyebrow and eyeball
2 gray scale valley points occur, eyeball is again in the lower section of eyebrow simultaneously, thus utilizes the level of gray scale
Coordinate determines eye center coordinate in vertical direction;According to the eyebrow obtained and eye center
Coordinate again extract the window comprising only eyes, owing to pupil is more black and the horizontal edge of eye socket
Obvious, in eyes window, the upright projection of gray scale and the upright projection of horizontal edge determine eyes
The horizontal coordinate at center;
Eye normalization is calibrated;
PCA template matching is accurately positioned eyes;
Measurement Relation extraction nose window according to face;
Window inner projection determines nose position;
Face information is expressed by utilizing discrete cosine transform and PCA template matching method to extract
The local feature that ability is strong, this local feature includes eyes, nose and face, utilizes people simultaneously
Face identification Fisherface method and simple spectrum holes method extract the global feature of face, merge
Local feature and global feature.
The step stating face by characteristic vector is as follows:
Positioning feature point algorithm is utilized to obtain the positional information of human face, according to the structure of face
Feature splits each organic region;Wherein, during eye areas is centrally located at two lines of centres
At Dian, size be 1.6de × 0.5de, de be two distances between centers after naturalization;Nasal area
Height sized by be 0.6de × 0.5de;
If I (x, y), Ic (x, y) and In (x, y) be respectively facial image, eye areas image and
Nasal area image, extracts each image information with DCT respectively:
Xh=Reshape (F (I), nh)
Xe=Reshape (F (Ie), ne)
Xn=Reshape (F (In), nn)
Wherein, Xh, Xe and Xn are respectively the DCT of facial image, eye areas and nasal area
Feature, (A, function n) is the upper left n × n submatrix of extraction two-dimensional matrix A to function Reshape
And this submatrix is converted to a n2 dimensional vector;Use series connection method, by vector Xh, Xe
Fusion feature vector Y0:Y0=(XhT, XeT, XnT) T is formed with Xn series connection;
Face assemblage characteristic vector Y:Y=(Y0-μ)/σ is obtained after removing mean normalization;
In formula, the mean vector of μ=E (Y0) training sample fusion feature;E () is mathematic expectaion letter
Number, σ is corresponding variance vectors.
Step S5: face characteristic data and the face characteristic in key population data base that will obtain
Data are compared;
Described step S5 is by special with the face in key population data base for the face characteristic data obtained
Levying the detailed description of the invention that data compare can be: will obtain the eigenvalue of optimum facial image
Generate after in key population data base, key population standard certificate photo eigenvalue is identified comparison
Similarity, similarity exceedes pre-set threshold value, then extract all personnel higher than predetermined threshold value
Image, and according to similarity height sequence, and certain concrete personnel corresponding.
In actual applications, time in described comparison process, specifically by described optimum facial image
Eigenvalue compare with the face characteristic in described blacklist.
In actual applications, time in described comparison process, specifically by described optimum facial image
Eigenvalue compare with the face characteristic in described white list.
Step S6: if discovery has the personnel of high similarity with personnel in key population data base,
Send information.
After described step S5 has performed, sort and key population can be obtained by similarity height
The comparing result that in data base, in personnel, concrete personnel are similar, then sends out according to this comparing result
Go out information.
In actual applications, judge according to comparing result, if finding and key population data
In storehouse, in personnel, blacklist has the personnel of high similarity, then indicate in monitored picture,
Show photo and the associated personal information of these personnel, send the information of specific rank.
In actual applications, judge according to comparing result, if finding and key population data
In storehouse, in personnel, blacklist has the personnel of high similarity, then indicate in video pictures,
Show photo and the associated personal information of this blacklist personnel, in conjunction with current monitoring place,
Send the alarm of different stage, point out on duty, patrol Security Personnel's further detailed tracking ratio
Right.
In actual applications, judge according to comparing result, if finding and key population data
In storehouse, in personnel, white list has the personnel of high similarity, then indicate in monitored picture,
Show photo and the associated personal information of these personnel, and send information and notify management personnel
Further confirm that.
Dynamic human face recognition methods described in the present embodiment can be in situation about coordinating without personnel
Under, automatically extract the optimum face in monitor video, warehouse-in preserve and with backstage key population number
Carry out real-time comparison according to storehouse, find the timely early warning of comparison result that similarity is higher, meet public affairs
The application demand of security protection video monitoring field management and control efficient to all kinds of personnel altogether.
A kind of dynamic human face recognition methods embodiment two
As it is shown on figure 3, a kind of dynamic human face recognition methods, described dynamic human face recognition methods bag
Include following steps:
Step S1: extract face characteristic data to set up key population data base;
Described step S1 extracts face characteristic data to set up being embodied as of key population data base
Mode can be: by obtaining related personnel's identity information with national public security system, household register system,
And personal information in key population data base, obtain all personnel by existing photograph image etc.
Face characteristic data, set up demographic data storehouse, set up for personnel in key population data base
Key population data base.
In actual applications, in described key population data base, personnel can be high-risk group, such as
The fugitive personnel in the whole nation, terrorist and previous conviction personnel etc., can be set to black by these high-risk group
List.
In actual applications, in described key population data base, personnel can be the crowd that passes through especially,
The management personnel in such as certain building, the current personnel of indult, People's Armed Police, public security etc., can
This kind of crowd is set to white list.
Step S2: real-time dynamic monitoring video is carried out moving body track;
Described step S2 carries out the specific embodiment party of moving body track to real-time dynamic monitoring video
Formula can be: by observing real-time dynamic monitoring video, the moving object to occurring in video is carried out
Follow the tracks of.
In actual applications, described real-time dynamic monitoring video is by DSP video camera or network
IP high-definition camera is monitored production, can take the photograph at described DSP video camera or network IP high definition
Camera performs step S4.
Step S3: whether there is face in the moving object of detecting and tracking and be ranked up;
Whether the moving object of described step S3 detecting and tracking exists face the tool being ranked up
Body embodiment can be: whether there is face in the moving object of detecting and tracking, and according to face
Movement locus and contour feature, by great amount of samples training obtain human-face detector, according to
Face eyes opening and closing situation, direction of visual lines and facial orientation carry out optimum facial image sequence.
In actual applications, optimum face can be with China second-generation identity card certificate photo as standard, just
Face is amimia, uniform light, without fuzzy, and the facial image of the eyeball that keeps one's eyes open.
In actual applications, the identification of optimum face can include following manner:
Facial feature localization----judge expression;
Attitude Algorithm----judge facial orientation;
Illumination algorithm----judge that luminosity is the most uniform, it is high light or backlight;
Fuzziness judges---whether image obscures.In actual applications, in the motion of detecting and tracking
When whether object exists face, move by following the tracks of target object, detect whether it is personnel,
Then proceed to detect the face whether these personnel occur in video, the most just face is tracked
Judge optimum face.
In actual applications, in the moving object of detecting and tracking, whether there is face and arrange
During sequence, can automatically lock picture occur everyone, the most anti-for the people hovered in picture
Multiple identification, at utmost reduces interference.
Step S4: extract face characteristic data inputting face characteristic number according to optimum facial image
According to storehouse;
Described step S4 extracts face characteristic data inputting face characteristic according to optimum facial image
Data base specifically can comprise the following steps that
Follow the tracks of optimum facial image;
Use Inline Function and fixed-point calculation that the facial image intercepted is carried out positioning feature point and spy
Value indicative is extracted;
The face characteristic value typing facial feature database that will extract.
In actual applications, according to optimum facial image and extract face characteristic data and comprise the steps that
Reducing broad image, attitude is corrected;
Carry out eigenvalue location, position including face, facial feature localization etc.;
Behind location, picture is normalized;
Calculate eigenvalue according to the picture after processing, specifically can pass through geometry location, use small echo
The modes such as conversion calculate eigenvalue.
In actual applications, eigenvalue comprises the steps that geometrical relationship;Color;Specific characteristic, than
Such as the nevus of face, birthmark etc..
In actual applications, the detailed description of the invention following the tracks of optimum facial image can be: to optimum
Facial image carries out Image semantic classification;Pretreated image is carried out Face datection;To detection
The face gone out carries out algorithm keeps track.
In actual applications, described employing Inline Function and the fixed-point calculation facial image to intercepting
Carrying out positioning feature point and characteristics extraction is high at DSP video camera or network IP when being embodied as
Clear video camera performs, in wherein Inline Function is the hardware that described video camera chips is corresponding
Connection function.
In actual applications, the described face characteristic value typing facial feature database that will extract
Before execution, described video camera needs the face characteristic value extracted and other face information data
It is transferred to server with typing facial feature database.
In actual applications, described employing Inline Function and the fixed-point calculation facial image to intercepting
The process that implements carrying out positioning feature point and characteristics extraction can be as follows:
A kind of facial modeling combined based on geometric projection and template matching is used to calculate
Method;First sciagraphy coarse positioning eye position is used;Then in this result, use PCA template
Matching method is accurately positioned;Finally according to the position location of eyes, use the 2 of sciagraphy location nose
Individual angle point and nose.
The key step of feature extraction is as follows:
Measurement Relation extraction eyebrow according to face and eyes window;
To eyebrow and eyes window inner projection coarse positioning eye position;The eyebrow and the eyes window that obtain are
Rectangle frame, if rectangle frame left upper apex coordinate and the coordinate of bottom right vertex, closes according to projection function
System, in calculation block any point in the horizontal direction with the average gray value in vertical direction, eyebrow
Hair and eyeball are black compared with other location comparisons, and gray value is in the horizontal direction at eyebrow and eyeball
2 gray scale valley points occur, eyeball is again in the lower section of eyebrow simultaneously, thus utilizes the level of gray scale
Coordinate determines eye center coordinate in vertical direction;According to the eyebrow obtained and eye center
Coordinate again extract the window comprising only eyes, owing to pupil is more black and the horizontal edge of eye socket
Obvious, in eyes window, the upright projection of gray scale and the upright projection of horizontal edge determine eyes
The horizontal coordinate at center;
Eye normalization is calibrated;
PCA template matching is accurately positioned eyes;
Measurement Relation extraction nose window according to face;
Window inner projection determines nose position;
Face information is expressed by utilizing discrete cosine transform and PCA template matching method to extract
The local feature that ability is strong, this local feature includes eyes, nose and face, utilizes people simultaneously
Face identification Fisherface method and simple spectrum holes method extract the global feature of face, merge
Local feature and global feature.
The step stating face by characteristic vector is as follows:
Positioning feature point algorithm is utilized to obtain the positional information of human face, according to the structure of face
Feature splits each organic region;Wherein, during eye areas is centrally located at two lines of centres
At Dian, size be 1.6de × 0.5de, de be two distances between centers after naturalization;Nasal area
Height sized by be 0.6de × 0.5de;
If I (x, y), Ic (x, y) and In (x, y) be respectively facial image, eye areas image and
Nasal area image, extracts each image information with DCT respectively:
Xh=Reshape (F (I), nh)
Xe=Reshape (F (Ie), ne)
Xn=Reshape (F (In), nn)
Wherein, Xh, Xe and Xn are respectively the DCT of facial image, eye areas and nasal area
Feature, (A, function n) is the upper left n × n submatrix of extraction two-dimensional matrix A to function Reshape
And this submatrix is converted to a n2 dimensional vector;Use series connection method, by vector Xh, Xe
Fusion feature vector Y0:Y0=(XhT, XeT, XnT) T is formed with Xn series connection;
Face assemblage characteristic vector Y:Y=(Y0-μ)/σ is obtained after removing mean normalization;
In formula, the mean vector of μ=E (Y0) training sample fusion feature;E () is mathematic expectaion letter
Number, σ is corresponding variance vectors.
Step S5: face characteristic data and the face characteristic in key population data base that will obtain
Data are compared;
Described step S5 is by special with the face in key population data base for the face characteristic data obtained
Levying the detailed description of the invention that data compare can be: will obtain the eigenvalue of optimum facial image
Generate after in key population data base, key population standard certificate photo eigenvalue is identified comparison
Similarity, similarity exceedes pre-set threshold value, then extract all personnel higher than predetermined threshold value
Image, and according to similarity height sequence, and certain concrete personnel corresponding.
In actual applications, time in described comparison process, specifically by described optimum facial image
Eigenvalue compare with the face characteristic in described blacklist.
In actual applications, time in described comparison process, specifically by described optimum facial image
Eigenvalue compare with the face characteristic in described white list.
Step S6: if discovery has the personnel of high similarity with personnel in key population data base,
Send information.
After described step S5 has performed, sort and key population can be obtained by similarity height
The comparing result that in data base, in personnel, concrete personnel are similar, then sends out according to this comparing result
Go out information.
In actual applications, judge according to comparing result, if finding and key population data
In storehouse, in personnel, blacklist has the personnel of high similarity, then indicate in monitored picture,
Show photo and the associated personal information of these personnel, send the information of specific rank.
In actual applications, judge according to comparing result, if finding and key population data
In storehouse, in personnel, blacklist has the personnel of high similarity, then indicate in video pictures,
Show photo and the associated personal information of this blacklist personnel, in conjunction with current monitoring place,
Send the alarm of different stage, point out on duty, patrol Security Personnel's further detailed tracking ratio
Right.
In actual applications, judge according to comparing result, if finding and key population data
In storehouse, in personnel, white list has the personnel of high similarity, then indicate in monitored picture,
Show photo and the associated personal information of these personnel, and send information and notify management personnel
Further confirm that.
Step S7: in monitor video, face is tracked according to eigenvalue;
Described step S7 is according to being embodied as that face is tracked in monitor video by eigenvalue
Mode can be: the face features value trace face extracted according to step S4, was following the tracks of
Needing in journey to be continually changing eigenvalue according to face characteristic, such as front just transfers lateral feature value to
Can be along with change.Such that it is able to obtain people be tracked meeting some eigenvalue.
In actual applications, built by public safety face identification system, it is achieved personnel tracking
And alarm mode is gathered information (face character) by artificial collection of information afterwards to dynamic realtime
Transformation, fully grasp hotel, Internet bar, bayonet socket, etc. omnibearing multidate information, public
Peace relevant departments build recognition of face comparison cloud computing platform, carry out " the most pre-to emphasis personnel
Evidence obtaining, follow-up tracking in police, thing " whole process supervision, it is achieved " allow each photographic head become
24 hours round-the-clock electronic polices ", embody public security science and technology and warn by force, improve existing society video
Science and high efficiency, make public security in time, accurately, dynamically grasp local safety situation,
Realize the new effect in advance prevented.
Dynamic human face recognition methods described in the present embodiment can be in situation about coordinating without personnel
Under, automatically extract the optimum face in monitor video, warehouse-in preserve and with backstage key population number
Carry out real-time comparison according to storehouse, find the timely early warning of comparison result that similarity is higher, meet public affairs
The application demand of security protection video monitoring field management and control efficient to all kinds of personnel altogether.
A kind of dynamic human face recognition system embodiment
As in figure 2 it is shown, a kind of dynamic human face recognition system, described dynamic human face recognition system bag
Include:
DBM 1, is used for extracting face characteristic data to set up key population data base;
Motion tracking module 2, for carrying out moving body track to real-time dynamic monitoring video;
Whether image collection module 3, exist face in the moving object of detecting and tracking and carry out
Sequence;
Characteristic extracting module 4, for extracting face characteristic data inputting according to optimum facial image
Facial feature database;
Feature comparing module 5, in the face characteristic data that will obtain and key population data base
Face characteristic data compare;
Reminding module 6, for finding have high similarity with personnel in key population data base
Information is sent during personnel.
In actual applications, described image collection module 3 includes: optimum face judge module,
For the movement locus according to face and contour feature, obtain face by great amount of samples training
Detector, carries out optimum face according to face eyes opening and closing situation, direction of visual lines and facial orientation
Image sorts.
In actual applications, described dynamic human face recognition system may also include camera driver module,
Described camera driver module is to write driving according to each producer SDK, can receive and decode
Camera video.
Wherein, described face characteristic data inputting face characteristic is extracted according to optimum facial image
Data base includes:
Follow the tracks of optimum facial image;
Use Inline Function and fixed-point calculation that the facial image intercepted is carried out positioning feature point and spy
Value indicative is extracted;
The face characteristic value typing facial feature database that will extract.
The advantage that the present invention implements: dynamic human face recognition methods of the present invention includes following step
Rapid: to extract face characteristic data to set up key population data base;To real-time dynamic monitoring video
Carry out moving body track;Whether the moving object of detecting and tracking exists face and is ranked up;
Face characteristic data inputting facial feature database is extracted according to optimum facial image;To obtain
Face characteristic data compare with the face characteristic data in key population data base;If sending out
The personnel now with personnel in key population data base with high similarity then send information, can
, in the case of coordinating without personnel, to automatically extract the optimum face in monitor video, warehouse-in
Preserve and carry out real-time comparison with backstage key population data base, finding the comparison that similarity is higher
Result is pointed out in time, meets the monitoring field management and control efficient to all kinds of personnel of public security protection video
Application demand.
The above, the only detailed description of the invention of the present invention, but protection scope of the present invention is also
Being not limited to this, any those skilled in the art is at technology model disclosed by the invention
In enclosing, the change that can readily occur in or replacement, all should contain within protection scope of the present invention.
Therefore, protection scope of the present invention should be as the criterion with described scope of the claims.
Claims (10)
1. a dynamic human face recognition methods, it is characterised in that described dynamic human face recognition methods
Comprise the following steps:
Extract face characteristic data to set up key population data base;
Real-time dynamic monitoring video is carried out moving body track;
Whether the moving object of detecting and tracking exists face and is ranked up;
Face characteristic data inputting facial feature database is extracted according to optimum facial image;
The face characteristic data of acquisition are carried out with the face characteristic data in key population data base
Comparison;
If discovery has the personnel of high similarity with personnel in key population data base, send prompting
Information.
Dynamic human face recognition methods the most according to claim 1, it is characterised in that described
Whether the moving object of detecting and tracking exists face and is ranked up concretely: detecting and tracking
Moving object in whether there is face, and according to the movement locus of face and contour feature,
Human-face detector is obtained, according to face eyes opening and closing situation, sight line side by great amount of samples training
Optimum facial image sequence is carried out to facial orientation.
Dynamic human face recognition methods the most according to claim 1, it is characterised in that described
Dynamic human face recognition methods is further comprising the steps of: if finding and personnel in key population data base
The personnel with high similarity are identified in same monitoring region repeatedly, then can assert these personnel
Hover in this region, and send information.
Dynamic human face recognition methods the most according to claim 1, it is characterised in that described
Dynamic human face recognition methods is further comprising the steps of: if find several with key population data base in
Personnel have the personnel of high similarity and are identified in same monitoring region, then send information.
Dynamic human face recognition methods the most according to claim 1, it is characterised in that described
According to optimum facial image and extract face characteristic data and include:
Reducing broad image, attitude is corrected;
Carry out eigenvalue location;
Behind location, picture is normalized;
Eigenvalue is calculated according to the picture after processing.
6., according to the dynamic human face recognition methods one of claim 1 to 5 Suo Shu, its feature exists
In, described extract face characteristic data inputting facial feature database bag according to optimum facial image
Include:
Follow the tracks of optimum facial image;
Use Inline Function and fixed-point calculation that the facial image intercepted is carried out positioning feature point and spy
Value indicative is extracted;
The face characteristic value typing facial feature database that will extract.
Dynamic human face recognition methods the most according to claim 6, it is characterised in that described
If discovery has the personnel of high similarity with personnel in key population data base, send information
Concretely: if discovery and personnel in key population data base have the personnel of high similarity, then
Monitored picture indicates, shows photo and the associated personal information of these personnel, send
The information of specific rank.
8. a dynamic human face recognition system, it is characterised in that described dynamic human face recognition system
Including:
DBM, is used for extracting face characteristic data to set up key population data base;
Motion tracking module, for carrying out moving body track to real-time dynamic monitoring video;
Whether image collection module, exist face in the moving object of detecting and tracking and carry out
Sequence;
Characteristic extracting module, for extracting face characteristic data inputting according to optimum facial image
Facial feature database;
Feature comparing module, in the face characteristic data that will obtain and key population data base
Face characteristic data compare;
Reminding module, for finding have high similarity with personnel in key population data base
Information is sent during personnel.
Dynamic human face recognition system the most according to claim 8, it is characterised in that described
Image collection module includes: optimum face judge module, for the movement locus according to face with
And contour feature, obtain human-face detector, according to face eyes opening and closing by great amount of samples training
Situation, direction of visual lines and facial orientation carry out optimum facial image sequence.
The most according to Claim 8, to the dynamic human face recognition system one of 9 described, its feature exists
In, described extract face characteristic data inputting facial feature database bag according to optimum facial image
Include:
Follow the tracks of optimum facial image;
Use Inline Function and fixed-point calculation that the facial image intercepted is carried out positioning feature point and spy
Value indicative is extracted;
The face characteristic value typing facial feature database that will extract.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510102708.2A CN106156688A (en) | 2015-03-10 | 2015-03-10 | A kind of dynamic human face recognition methods and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510102708.2A CN106156688A (en) | 2015-03-10 | 2015-03-10 | A kind of dynamic human face recognition methods and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106156688A true CN106156688A (en) | 2016-11-23 |
Family
ID=58063704
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510102708.2A Pending CN106156688A (en) | 2015-03-10 | 2015-03-10 | A kind of dynamic human face recognition methods and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106156688A (en) |
Cited By (34)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106650652A (en) * | 2016-12-14 | 2017-05-10 | 黄先开 | Trajectory tracking system and method based on face recognition technology |
CN106781236A (en) * | 2016-12-29 | 2017-05-31 | 东南大学 | A kind of based on person detecting and again remote alarming device and long-distance monitoring method of recognition methods |
CN106960203A (en) * | 2017-04-28 | 2017-07-18 | 北京搜狐新媒体信息技术有限公司 | A kind of facial feature tracking method and system |
CN107194357A (en) * | 2017-05-25 | 2017-09-22 | 湖南警察学院 | Face identification system and its detection method |
CN107451446A (en) * | 2017-07-18 | 2017-12-08 | 广东欧珀移动通信有限公司 | Solve lock control method and Related product |
CN107609508A (en) * | 2017-09-08 | 2018-01-19 | 深圳市金立通信设备有限公司 | A kind of face identification method, terminal and computer-readable recording medium |
CN108256459A (en) * | 2018-01-10 | 2018-07-06 | 北京博睿视科技有限责任公司 | Library algorithm is built in detector gate recognition of face and face based on multiple-camera fusion automatically |
CN108805140A (en) * | 2018-05-23 | 2018-11-13 | 国政通科技股份有限公司 | A kind of feature rapid extracting method and face identification system based on LBP |
CN109145714A (en) * | 2018-07-02 | 2019-01-04 | 深圳云感物联网科技有限公司 | A kind of identification of facial image carries out tracking and early warning method to emphasis control personnel |
CN109214300A (en) * | 2018-08-09 | 2019-01-15 | 成都智达万应科技有限公司 | A kind of suspect's method for early warning and its system based on recognition of face |
CN109214354A (en) * | 2018-09-26 | 2019-01-15 | 珠海横琴现联盛科技发展有限公司 | Dynamic human face retrieval method based on 3D camera |
CN109255303A (en) * | 2018-08-16 | 2019-01-22 | 南京西三艾电子***工程有限公司 | A kind of face identification system and method for the dynamic stream of people |
CN109389793A (en) * | 2018-10-16 | 2019-02-26 | 深圳众宝城贸易有限公司 | A kind of super anti-theft monitoring system of quotient |
CN109472247A (en) * | 2018-11-16 | 2019-03-15 | 西安电子科技大学 | Face identification method based on the non-formula of deep learning |
CN109492513A (en) * | 2018-05-24 | 2019-03-19 | 曜科智能科技(上海)有限公司 | The face space De-weight method of light field monitoring |
CN109522853A (en) * | 2018-11-22 | 2019-03-26 | 湖南众智君赢科技有限公司 | Face datection and searching method towards monitor video |
CN109614916A (en) * | 2018-12-06 | 2019-04-12 | 郑州铁路职业技术学院 | Station passenger status real time monitor early warning computer system based on recognition of face |
CN109801462A (en) * | 2019-01-21 | 2019-05-24 | 江西阳光安全设备集团有限公司 | The burglar alarm of intelligent mobile rack |
CN109815839A (en) * | 2018-12-29 | 2019-05-28 | 深圳云天励飞技术有限公司 | Hover personal identification method and Related product under micro services framework |
CN109858371A (en) * | 2018-12-29 | 2019-06-07 | 深圳云天励飞技术有限公司 | The method and device of recognition of face |
CN110032967A (en) * | 2019-04-10 | 2019-07-19 | 深圳市万睿智能科技有限公司 | A method of trailing personnel are identified under the current scene of more people |
CN110543868A (en) * | 2019-09-09 | 2019-12-06 | 福建省趋普物联科技有限公司 | Monitoring method and system based on face recognition and head and shoulder detection |
CN110728166A (en) * | 2018-07-16 | 2020-01-24 | 博博熊教育科技(中山)有限公司 | Intelligent monitoring integrated system for confirming track by face recognition |
CN110826371A (en) * | 2018-08-10 | 2020-02-21 | 京东数字科技控股有限公司 | Animal identification method, device, medium and electronic equipment |
CN110941993A (en) * | 2019-10-30 | 2020-03-31 | 东北大学 | Dynamic personnel classification and storage method based on face recognition |
CN111210375A (en) * | 2019-11-27 | 2020-05-29 | 重庆特斯联智慧科技股份有限公司 | Multi-functional portable wisdom security protection all-in-one |
CN111385440A (en) * | 2018-12-27 | 2020-07-07 | 芜湖潜思智能科技有限公司 | Monitoring camera with face recording and inquiring functions |
CN111597872A (en) * | 2020-03-27 | 2020-08-28 | 北京梦天门科技股份有限公司 | Health supervision law enforcement illegal medical practice face recognition method based on deep learning |
CN111860115A (en) * | 2020-06-03 | 2020-10-30 | 深圳市迈航信息技术有限公司 | Dynamic video portrait recognition method |
CN112597861A (en) * | 2020-12-16 | 2021-04-02 | 南京甄视智能科技有限公司 | Mixed-mode face recognition method and system supporting high-frequency use of buffer mechanism |
CN112784712A (en) * | 2021-01-08 | 2021-05-11 | 重庆创通联智物联网有限公司 | Missing child early warning implementation method and device based on real-time monitoring |
CN113095110A (en) * | 2019-12-23 | 2021-07-09 | 浙江宇视科技有限公司 | Method, device, medium and electronic equipment for dynamically warehousing face data |
TWI764425B (en) * | 2020-12-10 | 2022-05-11 | 鴻海精密工業股份有限公司 | Real time pedestrian statistical method based on face identification, and apparatus thereof |
US11403875B2 (en) | 2019-12-25 | 2022-08-02 | Askey Computer Corp. | Processing method of learning face recognition by artificial intelligence module |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101051385A (en) * | 2006-04-07 | 2007-10-10 | 欧姆龙株式会社 | Tracking method and device for special shooted objects and tracking method and device for aspect parts |
CN101174298A (en) * | 2006-11-02 | 2008-05-07 | 上海银晨智能识别科技有限公司 | Scattered-point high-volume face recognition system and recognizing method thereof |
CN101236599A (en) * | 2007-12-29 | 2008-08-06 | 浙江工业大学 | Human face recognition detection device based on multi- video camera information integration |
CN103310179A (en) * | 2012-03-06 | 2013-09-18 | 上海骏聿数码科技有限公司 | Method and system for optimal attitude detection based on face recognition technology |
-
2015
- 2015-03-10 CN CN201510102708.2A patent/CN106156688A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101051385A (en) * | 2006-04-07 | 2007-10-10 | 欧姆龙株式会社 | Tracking method and device for special shooted objects and tracking method and device for aspect parts |
CN101174298A (en) * | 2006-11-02 | 2008-05-07 | 上海银晨智能识别科技有限公司 | Scattered-point high-volume face recognition system and recognizing method thereof |
CN101236599A (en) * | 2007-12-29 | 2008-08-06 | 浙江工业大学 | Human face recognition detection device based on multi- video camera information integration |
CN103310179A (en) * | 2012-03-06 | 2013-09-18 | 上海骏聿数码科技有限公司 | Method and system for optimal attitude detection based on face recognition technology |
Cited By (47)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106650652A (en) * | 2016-12-14 | 2017-05-10 | 黄先开 | Trajectory tracking system and method based on face recognition technology |
CN106781236A (en) * | 2016-12-29 | 2017-05-31 | 东南大学 | A kind of based on person detecting and again remote alarming device and long-distance monitoring method of recognition methods |
CN106960203A (en) * | 2017-04-28 | 2017-07-18 | 北京搜狐新媒体信息技术有限公司 | A kind of facial feature tracking method and system |
CN107194357A (en) * | 2017-05-25 | 2017-09-22 | 湖南警察学院 | Face identification system and its detection method |
US11055547B2 (en) | 2017-07-18 | 2021-07-06 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Unlocking control method and related products |
CN107451446A (en) * | 2017-07-18 | 2017-12-08 | 广东欧珀移动通信有限公司 | Solve lock control method and Related product |
CN107451446B (en) * | 2017-07-18 | 2020-05-26 | Oppo广东移动通信有限公司 | Unlocking control method and related product |
CN107609508A (en) * | 2017-09-08 | 2018-01-19 | 深圳市金立通信设备有限公司 | A kind of face identification method, terminal and computer-readable recording medium |
CN108256459B (en) * | 2018-01-10 | 2021-08-24 | 北京博睿视科技有限责任公司 | Security check door face recognition and face automatic library building algorithm based on multi-camera fusion |
CN108256459A (en) * | 2018-01-10 | 2018-07-06 | 北京博睿视科技有限责任公司 | Library algorithm is built in detector gate recognition of face and face based on multiple-camera fusion automatically |
CN108805140B (en) * | 2018-05-23 | 2021-06-29 | 国政通科技股份有限公司 | LBP-based rapid feature extraction method and face recognition system |
CN108805140A (en) * | 2018-05-23 | 2018-11-13 | 国政通科技股份有限公司 | A kind of feature rapid extracting method and face identification system based on LBP |
CN109492513B (en) * | 2018-05-24 | 2023-04-07 | 曜科智能科技(上海)有限公司 | Face space duplication eliminating method for light field monitoring |
CN109492513A (en) * | 2018-05-24 | 2019-03-19 | 曜科智能科技(上海)有限公司 | The face space De-weight method of light field monitoring |
CN109145714A (en) * | 2018-07-02 | 2019-01-04 | 深圳云感物联网科技有限公司 | A kind of identification of facial image carries out tracking and early warning method to emphasis control personnel |
CN110728166A (en) * | 2018-07-16 | 2020-01-24 | 博博熊教育科技(中山)有限公司 | Intelligent monitoring integrated system for confirming track by face recognition |
CN109214300A (en) * | 2018-08-09 | 2019-01-15 | 成都智达万应科技有限公司 | A kind of suspect's method for early warning and its system based on recognition of face |
CN110826371A (en) * | 2018-08-10 | 2020-02-21 | 京东数字科技控股有限公司 | Animal identification method, device, medium and electronic equipment |
CN109255303A (en) * | 2018-08-16 | 2019-01-22 | 南京西三艾电子***工程有限公司 | A kind of face identification system and method for the dynamic stream of people |
CN109214354A (en) * | 2018-09-26 | 2019-01-15 | 珠海横琴现联盛科技发展有限公司 | Dynamic human face retrieval method based on 3D camera |
CN109214354B (en) * | 2018-09-26 | 2022-01-25 | 珠海横琴井通容智科技信息有限公司 | Dynamic face retrieval and identification method based on 3D camera |
CN109389793B (en) * | 2018-10-16 | 2021-05-28 | 南京溧水高新产业股权投资有限公司 | Superbusiness anti-theft monitoring system |
CN109389793A (en) * | 2018-10-16 | 2019-02-26 | 深圳众宝城贸易有限公司 | A kind of super anti-theft monitoring system of quotient |
CN109472247A (en) * | 2018-11-16 | 2019-03-15 | 西安电子科技大学 | Face identification method based on the non-formula of deep learning |
CN109472247B (en) * | 2018-11-16 | 2021-11-30 | 西安电子科技大学 | Face recognition method based on deep learning non-fit type |
CN109522853B (en) * | 2018-11-22 | 2019-11-19 | 湖南众智君赢科技有限公司 | Face datection and searching method towards monitor video |
CN109522853A (en) * | 2018-11-22 | 2019-03-26 | 湖南众智君赢科技有限公司 | Face datection and searching method towards monitor video |
CN109614916A (en) * | 2018-12-06 | 2019-04-12 | 郑州铁路职业技术学院 | Station passenger status real time monitor early warning computer system based on recognition of face |
CN111385440A (en) * | 2018-12-27 | 2020-07-07 | 芜湖潜思智能科技有限公司 | Monitoring camera with face recording and inquiring functions |
CN109815839A (en) * | 2018-12-29 | 2019-05-28 | 深圳云天励飞技术有限公司 | Hover personal identification method and Related product under micro services framework |
CN109858371A (en) * | 2018-12-29 | 2019-06-07 | 深圳云天励飞技术有限公司 | The method and device of recognition of face |
CN109801462A (en) * | 2019-01-21 | 2019-05-24 | 江西阳光安全设备集团有限公司 | The burglar alarm of intelligent mobile rack |
CN110032967A (en) * | 2019-04-10 | 2019-07-19 | 深圳市万睿智能科技有限公司 | A method of trailing personnel are identified under the current scene of more people |
CN110543868A (en) * | 2019-09-09 | 2019-12-06 | 福建省趋普物联科技有限公司 | Monitoring method and system based on face recognition and head and shoulder detection |
CN110941993A (en) * | 2019-10-30 | 2020-03-31 | 东北大学 | Dynamic personnel classification and storage method based on face recognition |
CN111210375A (en) * | 2019-11-27 | 2020-05-29 | 重庆特斯联智慧科技股份有限公司 | Multi-functional portable wisdom security protection all-in-one |
CN113095110B (en) * | 2019-12-23 | 2024-03-08 | 浙江宇视科技有限公司 | Method, device, medium and electronic equipment for dynamically warehousing face data |
CN113095110A (en) * | 2019-12-23 | 2021-07-09 | 浙江宇视科技有限公司 | Method, device, medium and electronic equipment for dynamically warehousing face data |
US11403875B2 (en) | 2019-12-25 | 2022-08-02 | Askey Computer Corp. | Processing method of learning face recognition by artificial intelligence module |
TWI783199B (en) * | 2019-12-25 | 2022-11-11 | 亞旭電腦股份有限公司 | Processing method of face recognition and electronic device |
CN111597872A (en) * | 2020-03-27 | 2020-08-28 | 北京梦天门科技股份有限公司 | Health supervision law enforcement illegal medical practice face recognition method based on deep learning |
CN111860115A (en) * | 2020-06-03 | 2020-10-30 | 深圳市迈航信息技术有限公司 | Dynamic video portrait recognition method |
TWI764425B (en) * | 2020-12-10 | 2022-05-11 | 鴻海精密工業股份有限公司 | Real time pedestrian statistical method based on face identification, and apparatus thereof |
CN112597861B (en) * | 2020-12-16 | 2022-03-18 | 南京甄视智能科技有限公司 | Mixed-mode face recognition method and system supporting high-frequency use of buffer mechanism |
CN112597861A (en) * | 2020-12-16 | 2021-04-02 | 南京甄视智能科技有限公司 | Mixed-mode face recognition method and system supporting high-frequency use of buffer mechanism |
CN112784712A (en) * | 2021-01-08 | 2021-05-11 | 重庆创通联智物联网有限公司 | Missing child early warning implementation method and device based on real-time monitoring |
CN112784712B (en) * | 2021-01-08 | 2023-08-18 | 重庆创通联智物联网有限公司 | Missing child early warning implementation method and device based on real-time monitoring |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106156688A (en) | A kind of dynamic human face recognition methods and system | |
CN106203274B (en) | Real-time pedestrian detection system and method in video monitoring | |
CN104091176B (en) | Portrait comparison application technology in video | |
Gowsikhaa et al. | Suspicious Human Activity Detection from Surveillance Videos. | |
CN105095829B (en) | A kind of face identification method and system | |
CN106128022B (en) | A kind of wisdom gold eyeball identification violent action alarm method | |
CN109819208A (en) | A kind of dense population security monitoring management method based on artificial intelligence dynamic monitoring | |
CN107392112A (en) | A kind of facial expression recognizing method and its intelligent lock system of application | |
CN109145742A (en) | A kind of pedestrian recognition method and system | |
CN105138954A (en) | Image automatic screening, query and identification system | |
CN103632132A (en) | Face detection and recognition method based on skin color segmentation and template matching | |
CN101226597A (en) | Method and system for recognizing nights pedestrian based on thermal infrared gait | |
CN102902986A (en) | Automatic gender identification system and method | |
CN104504408A (en) | Human face identification comparing method and system for realizing the method | |
CN110728252B (en) | Face detection method applied to regional personnel motion trail monitoring | |
CN105138967B (en) | Biopsy method and device based on human eye area active state | |
CN111353338B (en) | Energy efficiency improvement method based on business hall video monitoring | |
RU2315352C2 (en) | Method and system for automatically finding three-dimensional images | |
CN116310933A (en) | Cross-mirror tracking early warning method and system for rail transit intrusion scene personnel | |
Lee et al. | Distinction Between Real Faces and Photos by Analysis of Face Data. | |
Ryan et al. | Scene invariant crowd counting and crowd occupancy analysis | |
CN111368803A (en) | Face recognition method and system | |
Zhang et al. | Gender recognition based on fusion on face and gait information | |
Saranya et al. | Computer Vision on Identifying Persons under Real Time Surveillance using IOT | |
Mohana et al. | Human action Recognition using STIP Techniques |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161123 |