A kind of method for detecting human face and device
Technical field
The invention belongs to field of face identification, more particularly to a kind of method for detecting human face and device.
Background technology
Recognition of face is a kind of biological identification technology that the facial feature information based on people carries out identification.Use video camera
Or image or video flowing of the camera collection containing face, and automatic detect and track face, and then to detecting in the picture
Face carry out a series of correlation techniques of face, generally also referred to as Identification of Images, face recognition.
The face that existing face identification method is mainly prestored by the human face data more currently gathered with system
Data are identified result.Because existing method needs the cooperation of user to gather human face data, and the face to currently gathering
Data dependency is high, therefore when the human face data currently gathered changes, such as shaved beard, changed hair style, be more
Glasses, expression is even become, these human face datas changed are likely to cause comparison to fail, so as to cause discrimination mistake
It is low.
The content of the invention
The embodiments of the invention provide a kind of method for detecting human face and device, it is intended to solves existing method when identifying face
The problem of caused discrimination is too low.
The embodiment of the present invention is achieved in that a kind of method for detecting human face, and methods described comprises the steps:
Place and/or time when obtaining and storing the face data of multiple citizens and obtain the face data;
Obtain the face data of a specified citizen;
The face data of multiple citizens of the face data of the specified citizen with storing is matched;
The relevant information that matching degree reaches the citizen of predetermined threshold value is obtained, the relevant information of the citizen is included described in acquisition
The probability and/or time that the place of citizen's face data and/or citizen occur in place.
The another object of the embodiment of the present invention is to provide a kind of human face detection device, and described device includes:
Data capture unit, for obtain and store multiple citizens face data and obtain the face data when
Place and/or time;
Citizen's face data collecting unit is specified, for obtaining the face data of a specified citizen;
Matching unit, for by the face data of the specified citizen with storage multiple citizens face data carry out
Match somebody with somebody;
Output unit, the relevant information for reaching the citizen of predetermined threshold value for obtaining matching degree, the related letter of the citizen
The probability and/or time that breath includes obtaining the place of citizen's face data and/or citizen occurs in place.
In embodiments of the present invention, due to specially coordinating without user, therefore the face data obtained is closer to citizen's
Daily state, therefore it is favorably improved discrimination.In addition, if the match is successful, the citizen that the match is successful can be obtained and appear in it
The temporal information in his place, so as to rapidly from different sides recognize the personal information of the citizen and behavioural characteristic and
Trip rule.
Brief description of the drawings
Fig. 1 is a kind of flow chart for method for detecting human face that first embodiment of the invention provides;
Fig. 2 is a kind of structure chart for human face detection device that second embodiment of the invention provides;
Fig. 3 is the structure chart for another human face detection device that second embodiment of the invention provides.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
The embodiment of the present invention specially coordinates the face data for obtaining multiple citizens and corresponding place, time letter without user
The information such as breath, face data and corresponding place, time by storing multiple citizens, when the face for needing one citizen of matching
During data, directly the face data of the citizen is matched with the information stored.
The embodiments of the invention provide one kind:Method for detecting human face and device.
Methods described includes:Place when obtaining and storing the face data of multiple citizens and obtain the face data
And/or the time;
When resident's face data construction database is obtained, pass through the depth of view information acquired in camera, Neng Gouzhi
The picture of the road camera reaches area.By the way that the depth of field of the camera near the camera is combined, can be somebody's turn to do
Camera section reaches spatial dimension, so as to the three-dimensional reduction using this spatial dimension as abstract space, with the abstract three-dimensional
Big data spatial information of the data in space as people's space of planes position of structure.
In follow-up spatial match is searched, the spatial information can be made full use of, knows the space category for being matched object
Property with a certain space reach probability.
Obtain the face data of a specified citizen;
The face data of multiple citizens of the face data of the specified citizen with storing is matched;
The relevant information that matching degree reaches the citizen of predetermined threshold value is obtained, the relevant information of the citizen is included described in acquisition
The probability and/or time that the place of citizen's face data and/or citizen occur in place.
Described device includes:Data capture unit, for obtaining and storing the face data of multiple citizens and obtain institute
Place and/or time when stating face data;
Citizen's face data collecting unit is specified, for obtaining the face data of a specified citizen;
Matching unit, for by the face data of the specified citizen with storage multiple citizens face data carry out
Match somebody with somebody;
Output unit, the relevant information for reaching the citizen of predetermined threshold value for obtaining matching degree, the related letter of the citizen
The probability and/or time that breath includes obtaining the place of citizen's face data and/or citizen occurs in place.
In the embodiment of the present invention, due to specially coordinating without user, therefore obtain face data closer to citizen day
Normal state, therefore it is favorably improved discrimination.In addition, if the match is successful, the citizen that the match is successful can be obtained and appear in other
The temporal information in place, so as to recognize the personal information of the citizen and behavioural characteristic from different sides rapidly and go out
Professional etiquette is restrained.
In order to illustrate technical solutions according to the invention, illustrated below by specific embodiment.
Embodiment one:
Fig. 1 shows a kind of method for detecting human face that first embodiment of the invention provides, and details are as follows:
In step S101, place when obtaining and storing the face data of multiple citizens and obtain the face data
And/or the time.
In the step, specially coordinate without user, the photo or video data that can be captured by Face datection obtain citizen
Face data, place, time when by the face data of the citizen of acquisition and obtaining the face data of the citizen are stored in
In database.
Preferably, it is described obtain and store multiple citizens face data and obtain the face data when place
And/or the step of time, specifically includes:
The video data for the camera that A1, collection public place are set up.A2, to the face data in the video data of collection
Carry out cross-matched.A3, correspondingly store the face data that the match is successful and the place and/or the time that obtain the face data.
The public place of the step includes the places such as parking lot, megastore, both sides of the road, park.In this step, adopt
Collect the video data of the substantial amounts of citizen of public place accumulation, then the facial number that includes of video data in each place of cross-matched
According to finally storing corresponding data.Such as obtain parking lot first face data after, by the face data of first with
Multiple face datas that the places such as megastore, both sides of the road, park obtain are matched respectively, to judge whether first occurs
In places such as megastore, both sides of the road, parks, if the match is successful, it is assumed that first appears in park, then storing first simultaneously goes out
The information such as the place in parking lot and park, time now, otherwise, only store a first and appear in the letter such as the place in parking lot, time
Breath.Certainly, place here can also be specific to some camera in parking lot, park.
In step s 102, the face data of a specified citizen is obtained.
In the step, specified citizen refers to the citizen that needs match with the face data of database purchase, the face of the citizen
Portion's data carry out Face datection acquisition by the image of the citizen to acquisition.
In step s 103, by the face data of the specified citizen and the face data progress of multiple citizens of storage
Match somebody with somebody.
In the step, the face data of the citizen of storage is found out from database, and by the face of each citizen of lookup
Portion's data are compared with the face data of specified citizen, and the result of comparison is represented using matching degree.Two face datas more connect
Closely, matching degree is higher.
Preferably, the face data of multiple citizens of the face data by the specified citizen with storing is matched
The step of specifically include:
B1, the face data that technology searches multiple citizens of storage one by one is searched by big data.B2, pass through facial image
Pattern feature the face data of the specified citizen is matched with the face data of citizen searched.
The pattern feature of above-mentioned facial image include histogram feature, color characteristic, template characteristic, architectural feature and
Haar features.
In the step, information useful in the pattern feature of facial image is picked out, and calculated using Adaboost study
Method carries out Face datection using these features.The Adaboost is a kind of iterative algorithm, and its core concept is to be directed to same instruction
Practice collection and train different grader (Weak Classifier), then these weak classifier sets are got up, form one it is stronger final
Grader(Strong classifier).
In step S104, the relevant information that matching degree reaches the citizen of predetermined threshold value, the related letter of the citizen are obtained
The probability and/or time that breath includes obtaining the place of citizen's face data and/or citizen occurs in place.
In the step, judge to specify whether the matching degree of the face data and the face data of the citizen stored of citizen is more than
Or equal to predetermined threshold value, if so, the specified citizen of judgement is same people with the citizen matched, then pass through the mutual chain of Various database
Relation is connect, obtains the relevant information that matching degree is more than or equal to the citizen of predetermined threshold value, for example, obtaining citizen appears in some ground
The time of side, title, are appeared in somewhere(Camera)Probability, citizen continuously appear in it is general between certain several camera
Time continuity or spatial continuity that rate, citizen occur in some camera etc..In order to improve the accuracy of matching, can incite somebody to action
It is default
Preferably, when the face data that no acquisition matches with the face data of specified citizen, accordingly result is exported.
Such as:Do not find matching picture information it is default etc..
In embodiments of the present invention, specially coordinate without user the face data for obtaining multiple citizens and corresponding place,
The information such as temporal information, face data and corresponding place, time by storing multiple citizens, when need match a citizen
Face data when, directly by the face data of the citizen with store information matched.Due to specially coordinating without user,
Therefore the face data obtained therefore is favorably improved discrimination closer to the daily state of citizen.In addition, if the match is successful,
The temporal information that the citizen that the match is successful appears in other places can be then obtained, so as to recognize from different sides rapidly
The personal information of the citizen and behavioural characteristic and trip rule.
Preferably, in step S104, it is described obtain the citizen that matching degree reaches predetermined threshold value relevant information the step of it
After comprise the steps:
Corresponding information is pushed to the citizen for obtaining relevant information.
In the present embodiment, according to the relevant information of the citizen of acquisition, the daily behavior for analyzing citizen is accustomed to, and then according to
These behavioural habits push corresponding information, including corresponding merchandise news, information on services etc..In addition, the present embodiment can be with
Investigated into a case for public security, the data of the offer very multidimensional such as population purpose identification are true.
Embodiment two:
Fig. 2 shows a kind of structure for human face detection device that second embodiment of the invention provides, for convenience of description, only
Show the part related to the embodiment of the present invention.
The human face detection device can be used for whole by wired or wireless network connection server various information processings
End, such as mobile phone, pocket computer(Pocket Personal Computer, PPC), palm PC, computer, notes
This computer, personal digital assistant(Personal Digital Assistant, PDA)Deng can be operate at these information
The unit that software unit, hardware cell or software and hardware in reason terminal are combined, can also be integrated into as independent suspension member
In these information processing terminals or run in the application system of these information processing terminals, wherein:
Data capture unit 21, for obtain and store multiple citizens face data and obtain the face data when
Place and/or the time.
Photo or video data that the data capture unit 21 can be captured by Face datection obtain the face data of citizen.
Preferably, the data capture unit 21 includes:Video data acquiring module, face data cross-matched module with
And memory module.Wherein:
Video data acquiring module, the video data of the camera for gathering public place erection.
Face data cross-matched module, cross-matched is carried out for the face data in the video data to collection.
Memory module, for correspondingly store the face data that the match is successful and obtain the face data place and/or
Time.
Wherein, public place includes the places such as parking lot, megastore, both sides of the road, park.By being deposited to each place
Video data carry out logic connection, so as to be built into knowledge base, and big data excavation is carried out to the knowledge base, can be effective
Quickly its more more complete relevant information according to the image surface acquisition of information of people, so as to recognize trip of the citizen on space-time
Rule, and trip custom.
Citizen's face data collecting unit 22 is specified, for obtaining the face data of a specified citizen.
Matching unit 23, for the face data of the specified citizen and the face data of multiple citizens of storage to be carried out
Matching.
The matching unit 23 represents the similarity of two face datas using the height of matching degree.
Preferably, the matching unit 23 includes:Searching modul and face matching module.Wherein:
Searching modul, the face data of multiple citizens for searching storage one by one by big data lookup technology.
Face matching module, for the pattern feature by facial image by the face data of the specified citizen and lookup
The face data of citizen matched.
The pattern feature of above-mentioned facial image includes histogram feature, color characteristic, template characteristic, architectural feature and Haar
Feature.
Output unit 24, the relevant information for reaching the citizen of predetermined threshold value for obtaining matching degree, the correlation of the citizen
The probability and/or time that place and/or citizen of the information including acquisition citizen's face data occur in place.
The output unit 24 is used to obtain relevant information of the matching degree more than or equal to the citizen corresponding to predetermined threshold value.It is excellent
Selection of land, when the face data that no acquisition matches with the face data of specified citizen, export accordingly result.Such as:Do not find
It is default etc. with Tu Pian information.
In the embodiment of the present invention, due to specially coordinating collection face data without user, therefore the face data obtained is more
Close to the daily state of citizen, therefore it is favorably improved discrimination.In addition, if the match is successful, the public affairs that the match is successful can be obtained
The people appear in the temporal information in other places, so as to recognize the personal information and row of the citizen from different sides rapidly
Be characterized and go on a journey rule.
Fig. 3 shows the structure of another human face detection device provided in an embodiment of the present invention, the human face detection device bag
Include data capture unit 21, specify citizen's face data collecting unit 22, matching unit 23, output unit 24 and push unit
25.Wherein:
Push unit 25, for pushing corresponding information to the citizen for obtaining relevant information.
On the basis of the daily behavior for analyzing citizen is accustomed to, the push unit 25 pushes and analysis result phase to citizen
The information met, such as corresponding merchandise news, information on services etc..
In embodiments of the present invention, the video data of each camera in public place is gathered, and is known by Face datection
The face data of other citizen, the face data and corresponding place, temporal information for obtaining multiple citizens are specially coordinated without user,
The information such as face data and corresponding place, time by storing multiple citizens, when the facial number for needing one citizen of matching
According to when, directly by the face data of the citizen with store information matched.Due to specially coordinating without user, therefore obtain
Face data closer to the daily state of citizen, therefore be favorably improved discrimination.In addition, if the match is successful, can obtain
The citizen that the match is successful appears in the temporal information in other places, so as to recognize the citizen's from different sides rapidly
Personal information and behavioural characteristic and trip rule.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
All any modification, equivalent and improvement made within refreshing and principle etc., should be included in the scope of the protection.