CN101587485B - Face information automatic login method based on face recognition technology - Google Patents
Face information automatic login method based on face recognition technology Download PDFInfo
- Publication number
- CN101587485B CN101587485B CN2009100334764A CN200910033476A CN101587485B CN 101587485 B CN101587485 B CN 101587485B CN 2009100334764 A CN2009100334764 A CN 2009100334764A CN 200910033476 A CN200910033476 A CN 200910033476A CN 101587485 B CN101587485 B CN 101587485B
- Authority
- CN
- China
- Prior art keywords
- face
- people
- face information
- characteristic value
- recognition technology
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
The invention discloses a method for automatically logging in face information, which adopts a face detection and recognition technology, extracts face information with high frequency of appearance from collected video data or a plurality of different pictures through an automatic logging system, and automatically logs in, and comprises the following steps: (1) detecting a human face from a data source by adopting a human face detection technology, wherein the data source can be video data acquired by a camera or a picture; (2) in the off-line training stage, a face recognition technology is adopted to extract the characteristic value of the detected face and add the characteristic value into a characteristic value database; matching the faces in the characteristic value database by adopting a face recognition technology, and finding out face information with higher similarity in the database, wherein the face information is regarded as the face information of the same person; (3) in the online constraint stage, under objective conditions such as time and space constraints, the result of offline training is judged, and the offline training and the online constraint are combined to accurately judge the face information of the same person in the database; (4) and finding out the face information with the highest frequency of occurrence in the database, and automatically logging in the database.
Description
Technical field:
The present invention relates to a kind of people's face information automatic login method, belong to the face recognition technology field based on face recognition technology.
Background technology:
At present, face identification system is because its intuitive, convenience, untouchable, friendly, user's acceptance advantages of higher, is widely used aspect many in business administration, house safety, authentication or the like.
The job step of face identification system generally is divided into two ones: people's face information registration and people's face matching inquiry.Generally, in people's face information registration step, with people's face information of camera extracting destination object, perhaps photo of typing destination object adopts face recognition technology to extract people's face information characteristics value, and login people face information bank uses for matching inquiry; In people's face matching inquiry step, the data in the current facial image that collects and the people's face information bank are mated, seek the highest people's face information of similarity.
But there is the defective of following several respects in the method in common people's face information registration step:
(1) need destination object to login one by one, efficient is lower.Especially when destination object is more, this will be a very hard work.Such as, one has 1000 staff's business, needs people's face information of the typing staff, if adopt this traditional login mode, will expend a large amount of time and manpower.
(2) higher to the coherence request of ambient light.Studies show that, the feature difference of same individual under different light rays, even greater than at the feature difference between the different people under the same light, ambient light is very big to the influence that recognition performance produces, ambient light when matching inquiry is different with when login, the accuracy of inquiry has bigger decline, can't satisfy the requirement of application.
(3) owing to be typing one to one, every people's face information has only a piece of data in database, and eigenwert is comprehensive inadequately, also can reduce the accuracy when knowing matching inquiry.For example, if only logined people's face information of destination object in the database, when matching inquiry, if the distance of camera and people's face and when login are inconsistent, inconsistent when perhaps the angle of camera and people's face is with login, all can cause the accuracy of coupling to reduce.Studies show that the people's face information of the someone in the database is comprehensive more, many more, the accuracy of matching inquiry is high more.
(4) common people's face login system needs user's input interface, such as keyboard, when collecting people's face information of destination object, needs to key in and determines that information could be with people's face information registration to database.This traditional login mode for the smart machine of some no user input interfaces, such as robot, can't use, and therefore is unfavorable for general and expansion.
Summary of the invention:
The present invention is directed to the above-mentioned defective of traditional login mode, propose a kind of people's face information automatic login method based on face recognition technology.
Technical scheme of the present invention is achieved like this:
A kind of people's face information is the method for login automatically, adopts people's face detection and Identification technology, extracts the higher people's face information of the frequency of occurrences by automatic login system from the video data that collects or many different pictures, and logins automatically, and step is as follows:
(1) adopts human face detection tech, detect people's face from data source.Data source can be the video data that camera collection arrives, and also can be picture.Data source is abundant more, and people's face information is comprehensive more, and the accuracy during matching inquiry is high more;
(2) adopt face recognition technology, extract the eigenwert of detected people's face, add the characteristic value data storehouse, adopt face recognition technology, people's face in the characteristic value data storehouse is mated, find out the higher people's face information of similarity in the database, think people's face information of same individual, this stage is called off-line training;
(3) in some objective condition, under the constraint such as time, space, the result of off-line training is made judgement, this stage is called online constraint, and off-line training and online constraint combine, and accurately judge the people's face information of the same individual in the database;
(4) on the basis of the above work, find out people's face information that the frequency of occurrences is the highest in the database, it is signed in to database automatically.
A kind of people's face information is the system of login automatically, comprises the first-class video capture device of shooting or image data and automatic login device, and wherein automatic login device comprises people's face detection module, face recognition module, online constraints module and automatic login module.
The present invention logins automatically to the higher people of the frequency of occurrences in the facial image that collects.Improved the efficient of login, reduced the workload of login, because data source is abundanter, every people's face information has many piece of data in database, has improved the accuracy of identification.And be subjected to environmental influence less, can be widely used in aspects such as business administration, house safety, information security, have high economic benefit and social benefit.
Description of drawings:
Fig. 1 is an automatic login system structured flowchart of the present invention;
Fig. 2 is the automatic login device structured flowchart among Fig. 1;
Fig. 3 is a workflow diagram of the present invention.
Embodiment:
A kind of people's face information is the method for login automatically, adopts people's face detection and Identification technology, extracts the higher people's face information of the frequency of occurrences by automatic login system from the video data that collects or many different pictures, and logins automatically, and step is as follows:
(1) adopt human face detection tech, detect people's face from data source, data source can be the video data that camera collection arrives, and also can be picture, and data source is abundant more, and people's face information is comprehensive more, and the accuracy during matching inquiry is high more;
(2) adopt face recognition technology, extract the eigenwert of detected people's face, add the characteristic value data storehouse, adopt face recognition technology, people's face in the characteristic value data storehouse is mated, find out the higher people's face information of similarity in the database, think people's face information of same individual, this stage is called off-line training;
(3) in some objective condition, under the constraint such as time, space, the result of off-line training is made judgement, this stage is called online constraint; Off-line training and online constraint combine, and accurately judge the people's face information of the same individual in the database;
(4) on the basis of the above work, find out people's face information that the frequency of occurrences is the highest in the database, it is signed in to database automatically.
A kind of system of people's face information comprises the first-class video capture device of shooting or image data and automatic login device, and wherein automatic login device comprises people's face detection module, face recognition module, online constraints module and automatic login module.
Workflow of the present invention as shown in Figure 3.
Technical scheme of the present invention has been used following several committed step:
(1) adopt human face detection tech from data source, to extract people's face information.Adopt the eye location algorithm in the human face detection tech, at first the picture in the data source is treated to gray level image, former gray level image is being done edge gray scale reinforcement; In conjunction with the position of human eye decision criteria of setting up according to people's face geometric properties priori, in the process that segmentation threshold increases progressively, searching can be partitioned into the optimum segmentation threshold value of eyes eye piece then; The authenticity of coming the eyes that inspection goes out as symmetrical similarity at last with the bidimensional related coefficient, and utilize the gray integration projection of the binocular images vertical direction that finds, accurately locate pupil center.To coloured image, utilize the colour of skin to set up complexion model at the distribution character of YCbCr color space, find out area of skin color roughly, carry out greyscale transformation after, adopt said method to detect people's face again.
Such as, one has 50 staff's business, need do a Staffing System, can provide 200 different scenes (in meeting room, in the dining room, the gymnasium, camp or the like), different number (can be single photo, it also can be group picture, the personage not necessarily all is the company personnel based on the company personnel), different time is (such as the photo before 5 years, the photo before 2 years, present photo or the like) personage's photo is as data source, or the employee information in the continuous week that a plurality of monitoring cameras photograph in the company.
Adopt human face detection tech, detect the people's face information in video or the picture.Such as, in 200 photos of last example, detect 600 people's face information altogether, in these 600 facial images as training sample.
(2) off-line training---adopt face recognition technology, extract the face characteristic value, and mate.
At first adopt based on the face recognition algorithms of local feature analysis and optimization coupling and extract the face characteristic value, utilize neural network method to estimate some unique points that in identification people face, play an important role, pupil, canthus, place between the eyebrows, eyebrow pin, the corners of the mouth, nose utilize the multiple dimensioned feature of the local multiscale analysis feature extraction unique point of Gabor small echo afterwards.
After training sample extraction face characteristic value, it is added the sample characteristics storehouse.
In the sample characteristics storehouse, the eigenwert of each sample is mated, judge the deviation between each sample characteristics, if deviation, thinks then that similarity is higher less than some threshold values, judge that the higher people's face of similarity belongs to same individual.This stage is called off-line training.
Such as in last example, detected 600 facial images are extracted eigenwert, add the eigenwert storehouse.And the eigenwert of 600 people's faces in the eigenwert storehouse mated, two or many facial images that similarity is higher are thought same individual's facial image.Such as, in these 600 facial images, identifying wherein has the similarity of 10 facial images higher, then thinks these 10 facial images that facial image is same individual.
(3) carry out online constraint according to certain objective condition.
Off-line training only is to judge according to the similarity of face characteristic value, needs certain objective condition to carry out the result of off-line training is retrained and further judgement, comprising constraint conditions such as time, space, places.
Such as, in the higher facial image of 10 similarities of last example,, then can not be same individual if 2 or above appearing at simultaneously in same the photo are arranged.For another example, judging if go up in the higher facial image of 10 similarities of example, if at same shooting time, have 2 or above facial image to appear at different places or scene according to the shooting time of photo, then also can not be same individual.Under the constraint of objective condition, the result of off-line training is made judgement.
(4) frequency of occurrences is the highest same facial image is logined automatically.
Off-line training and online constraint combine, and accurately judge the people's face information of the same individual in the database.And on this basis, find out the highest same individual face information of the frequency of occurrences in the database, login automatically.Such as 600 facial images of last example, through off-line training and online constraint, to judge the facial image that a people is wherein arranged and occurred 15 times, the frequency of occurrences is the highest, thinks that then this person is the employee of company, then logs on as the company personnel with it automatically.Then, again the people's face information in the database is made judgement, find out the highest people's face picture of remaining frequency of occurrences, log on as the company personnel automatically.So circulation is lower than some threshold values up to the frequency that occurs, and thinks the underfrequency that this person occurs, and may not be the employee of company, so do not login, finishes.
The present invention logins automatically to the higher people of the frequency of occurrences in the facial image that collects, and has improved the efficient of login greatly, has saved manpower; Because data source is abundanter, every people's face information has many piece of data in database, has improved the accuracy of identification, and affected by environment less.Can be widely used in aspects such as business administration, house safety, information security, have high economic benefit and social benefit.
Claims (3)
1. one kind based on people's face information of face recognition technology method of login automatically, it is characterized in that: adopt people's face detection and Identification technology, from the video data that collects or many different pictures, extract the highest people's face information of the frequency of occurrences by automatic login system, and login automatically, step is as follows:
(1) adopt human face detection tech, detect people's face from data source, data source is the video data that camera collection arrives, or picture;
(2) off-line training step adopts face recognition technology, extracts the eigenwert of detected people's face, adds the characteristic value data storehouse; Adopt face recognition technology, the people's face in the characteristic value data storehouse is mated, find out the higher people's face information of similarity in the characteristic value data storehouse, think people's face information of same individual;
(3) the online constraint stage, under the constraint in objective condition time, space, the result of off-line training is made judgement, off-line training and online constraint combine, and accurately judge the people's face information of the same individual in the characteristic value data storehouse;
(4) find out people's face information that the frequency of occurrences is the highest in the characteristic value data storehouse, it is signed in to automatic login system automatically.
2. the method for logining automatically based on people's face information of face recognition technology according to claim 1, it is characterized in that: described automatic login system comprises data source and automatic login device.
3. the method for logining automatically based on people's face information of face recognition technology according to claim 2 is characterized in that: described automatic login device comprises people's face detection module, face recognition module, online constraints module and automatic login module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100334764A CN101587485B (en) | 2009-06-15 | 2009-06-15 | Face information automatic login method based on face recognition technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100334764A CN101587485B (en) | 2009-06-15 | 2009-06-15 | Face information automatic login method based on face recognition technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101587485A CN101587485A (en) | 2009-11-25 |
| CN101587485B true CN101587485B (en) | 2011-08-03 |
Family
ID=41371734
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009100334764A Active CN101587485B (en) | 2009-06-15 | 2009-06-15 | Face information automatic login method based on face recognition technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101587485B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102164113A (en) * | 2010-02-22 | 2011-08-24 | 深圳市联通万达科技有限公司 | Face recognition login method and system |
| CN102419819B (en) * | 2010-10-25 | 2014-10-08 | 深圳市中控生物识别技术有限公司 | Method and system for recognizing human face image |
| CN102637255A (en) * | 2011-02-12 | 2012-08-15 | 北京千橡网景科技发展有限公司 | Method and device for processing faces contained in images |
| CN103294986B (en) * | 2012-03-02 | 2019-04-09 | 汉王科技股份有限公司 | A kind of recognition methods of biological characteristic and device |
| CN102663359B (en) * | 2012-03-30 | 2014-04-09 | 博康智能网络科技股份有限公司 | Method and system for pedestrian retrieval based on internet of things |
| CN102750521A (en) * | 2012-06-05 | 2012-10-24 | 广东智华计算机科技有限公司 | Face-based identity recognition system |
| CN102867173B (en) * | 2012-08-28 | 2015-01-28 | 华南理工大学 | Human face recognition method and system thereof |
| CN102930261A (en) * | 2012-12-05 | 2013-02-13 | 上海市电力公司 | Face snapshot recognition method |
| CN104133917B (en) * | 2014-08-15 | 2018-08-10 | 百度在线网络技术（北京）有限公司 | The classification storage method and device of photo |
| CN108205620B (en) * | 2016-12-20 | 2021-07-02 | 航天信息股份有限公司 | Method for automatically logging in tax self-service terminal and tax self-service terminal |
| CN106897460A (en) * | 2017-03-14 | 2017-06-27 | 华平智慧信息技术（深圳）有限公司 | The method and device of data classification in safety monitoring |
| CN107302492A (en) * | 2017-06-28 | 2017-10-27 | 歌尔科技有限公司 | Friend-making requesting method, server, client terminal device and the system of social software |
| CN108734144A (en) * | 2018-05-28 | 2018-11-02 | 北京文香信息技术有限公司 | A kind of speaker's identity identifying method based on recognition of face |
| CN110796838B (en) * | 2019-12-03 | 2023-06-09 | 吉林大学 | Automatic positioning and recognition system for facial expression of driver |
| CN115953823B (en) * | 2023-03-13 | 2023-05-16 | 成都运荔枝科技有限公司 | Face recognition method based on big data |
-
2009
- 2009-06-15 CN CN2009100334764A patent/CN101587485B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN101587485A (en) | 2009-11-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101587485B (en) | Face information automatic login method based on face recognition technology | |
| CN102902959B (en) | Face recognition method and system for storing identification photo based on second-generation identity card | |
| WO2021036436A1 (en) | Facial recognition method and apparatus | |
| US9025864B2 (en) | Image clustering using a personal clothing model | |
| CN105354527A (en) | Negative expression recognizing and encouraging system | |
| JP5959093B2 (en) | People search system | |
| CN105825176A (en) | Identification method based on multi-mode non-contact identity characteristics | |
| CN105184238A (en) | Human face recognition method and system | |
| CN107230267B (en) | Intelligence In Baogang Kindergarten based on face recognition algorithms is registered method | |
| CN108009482A (en) | One kind improves recognition of face efficiency method | |
| JP2007317062A (en) | Person recognition apparatus and method | |
| CN109359625A (en) | The method and system of customer identification is judged based on head and shoulder detection and face recognition technology | |
| CN109145742A (en) | A kind of pedestrian recognition method and system | |
| US20070183634A1 (en) | Auto Individualization process based on a facial biometric anonymous ID Assignment | |
| CN105335691A (en) | Smiling face identification and encouragement system | |
| CN104091176A (en) | Technology for applying figure and head portrait comparison to videos | |
| Xue et al. | Clothing change aware person identification | |
| CN105975938A (en) | Smart community manager service system with dynamic face identification function | |
| CN106446753A (en) | Negative expression identifying and encouraging system | |
| CN109492538A (en) | Intelligent boarding system, method and readable storage medium storing program for executing based on face recognition technology | |
| US20120269389A1 (en) | Information processing apparatus, method of controlling information processing apparatus, and storage medium | |
| Shrivastava et al. | Conceptual model for proficient automated attendance system based on face recognition and gender classification using Haar-Cascade, LBPH algorithm along with LDA model | |
| CN204143555U (en) | The Certificate of House Property printing terminal of identification self-aided terminal and correspondence | |
| CN113505717B (en) | Online passing system based on face and facial feature recognition technology | |
| Tee et al. | Facial recognition using enhanced facial features k-nearest neighbor (k-NN) for attendance system |
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: 20151119 Address after: 214131 Jiangsu province Wuxi District Zhenze Road No. 18 Wuxi Software Park whale block A No. 6 Patentee after: Wuxi Le Chi Technology Co. Ltd. Address before: 214028 D building C202-1, creative innovation industrial park, 5, Wuxi New District, Jiangsu, Xinhua Road Patentee before: Wuxi Venpoo Technology Co., Ltd. |
|
| DD01 | Delivery of document by public notice |
Addressee: Li Qing Document name: Notification of Approving Refund |
|
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 214131 room 1601, block D, swan tower, Wuxi Software Park, No. 111, Linghu Avenue, Xinwu District, Wuxi City, Jiangsu Province Patentee after: Wuxi Leqi Technology Co.,Ltd. Address before: No. 6, tower a, whale tower, Wuxi Software Park, No. 18, Zhenze Road, Wuxi New Area, Jiangsu 214131 Patentee before: Wuxi Le Chi Technology Co.,Ltd. |