CN106980844A - A kind of character relation digging system and method based on face identification system - Google Patents
A kind of character relation digging system and method based on face identification system Download PDFInfo
- Publication number
- CN106980844A CN106980844A CN201710220835.1A CN201710220835A CN106980844A CN 106980844 A CN106980844 A CN 106980844A CN 201710220835 A CN201710220835 A CN 201710220835A CN 106980844 A CN106980844 A CN 106980844A
- Authority
- CN
- China
- Prior art keywords
- face
- record
- class
- vector
- characteristic vector
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a kind of character relation digging system based on face identification system and method, system includes video acquisition module, characteristic extracting module, data memory module, data processing module, the methods such as distance matching and cluster of the present invention using characteristic vector, automatic mining obtains the character relation between face picture, there is provided one and be queried list and relevant picture of the people by correlation degree ranking, improve the efficiency of public security technical search, effectively the workload of reduction personnel query, improves the accuracy that character relation is excavated.
Description
Technical field
The present invention relates to technical field of face recognition, and in particular to one kind carries out character relation digging based on face recognition technology
The system and method for pick.
Background technology
The typical technology that recognition of face is recognized as biometric identity, due to cooperating with one's own initiative for individual need not be detected, closely
In man-machine interaction over year, security protection, authentication, amusement, and got a lot of applications in terms of Medical nursing.Face recognition technology
Including：Face datection, feature extraction and characteristic matching and classification.The method of Face datection has：HARR is scanned, HOG scannings,
ADABOOT learns, deep learning CNN object detections etc..The method of feature extraction has：The intrinsic faces of PCA, deep learning CNN features
Extract etc..Characteristic matching and classification include：1-NN, k-NN and SVM.By various Face datections above-mentioned, feature extraction and
The method of characteristic matching is organically combined, it is possible to obtain face recognition technology general at present.
In smart city, in security protection and public security technical search, recognition of face is conventional artificial intelligence technology means.We
It is generally necessary to input a photo for being queried people in identifying system, be queried people by system return occurs on which camera
Cross, and the time point occurred and picture at that time, video segment.During public security technical search is carried out, we also need
Analyze be queried people appearance picture, from these pictures at lookup be queried people and meanwhile occur other with being queried people
The who object of dependency relation is there may be further to analyze.
, it is necessary to manually analyze Query Result in existing security protection public security deployment face system, pass through system
The picture of return, manually irises out the who object that other in picture there may be dependency relation with being queried people, then carries out phase
Close relation excavation and then input system carries out secondary inquiry.The excavation of existing character relation, can only not only work by artificial
Amount is big, wastes time and energy, and is easy to ignore relevant information, and due to the difference of experience, what different criminal detectives excavated
Great difference just occurs in character relation, and then can bring extra difficulty to criminal investigation work.Therefore it is badly in need of a kind of personage to close
It is the system and method for automatic mining, on the one hand reduces the workload of personnel query, improves the accuracy that character relation is excavated, separately
On the one hand it can eliminate due to the otherness that the character relation that inquirer's experience different band is come is excavated.
The content of the invention
For problem of the prior art, the present invention propose a kind of character relation digging system based on face identification system and
Method, using characteristic matching and the method for cluster, automatic mining obtain character relation between face picture there is provided one and
List and relevant picture of the people by correlation degree ranking are queried, the efficiency of public security technical search is improved.
The present invention is for the technical scheme that is used of solution above-mentioned technical problem：
The present invention provides a kind of character relation digging system based on face identification system, including video acquisition module, special
Levy extraction module, data memory module, data processing module；
The video acquisition module is the video acquisition device in face identification system, is generally laid in designated area
Camera, in real time gather specified range in video frame images；
The characteristic extracting module, for carrying out feature respectively to the multiple different faces occurred in each video frame images
Extract, the corresponding face characteristic vector of every face of generation simultaneously outputs this to data memory module；For being looked into input
The face for asking people carries out feature extraction, and generation is queried the corresponding face characteristic vector of people；
The data memory module is a face information database, the every record correspondence frame of video figure stored in database
The corresponding face information of each personage occurred as in, described face information includes face characteristic vector, face picture, people
The true picture that face occurs, video-frequency band and the unique frame number of total system that face occurs；
The data processing module, for every record to being queried in the characteristic vector and face information database of people
Characteristic vector enter row distance matching, return to qualified face record；For being gathered to qualified face record
Class processing, generates and records the face information table constituted by multigroup class；For according in every group of class record record object number, it is right
Multigroup class record is ranked up and exported.
A kind of character relation method for digging based on face identification system, comprises the following steps：
S1, obtains the video frame images sequence of video acquisition system collection in real time, multiple to what is occurred in each two field picture
Face picture carries out feature extraction respectively, the corresponding characteristic vector of every face of generation, and by this feature vector and corresponding people
The face information database of face information storage in the server；The face information of face information database purchase at least includes following
Content：Face picture, the frame picture that face occurs, video-frequency band and the unique number of frames of total system that face occurs.
S2, input is queried human face photo, and the face part on photo is cut out by detection, corrects and ajusts face,
Then feature extraction is carried out, obtains being queried the characteristic vector of people；
S3, the corresponding characteristic vector of every face picture on face information database and the characteristic vector being queried are carried out
Distance matching, if returning to face picture correspondence video frame number if less than predetermined threshold；
S4, is inquired about in face information database, obtains and remove with being looked into successively according to the step S3 frame numbers returned
The distance of characteristic vector of people is ask less than other all frame number identical faces records beyond the face record of predetermined threshold value,
And clustering processing is carried out, the class record sheet being made up of the corresponding class record of different target personage is generated, while calculating every group of class note
The central feature vector of record；
S5, is recorded for every group of class, every center for recording corresponding characteristic vector and this group of class record is calculated respectively special
The distance of vector is levied, the record that distance is more than predetermined threshold value is deleted；
S6, record according to included in every group of class record number according to descending sort and export.
The beneficial effects of the invention are as follows：
Between the face picture that the present invention can be obtained with automatic mining character relation is there is provided one and is queried people by pass
The list of connection degree ranking and relevant picture.Compared with current hand digging means, this significantly facilitates and improved public security technology
The efficiency of scouting.
Brief description of the drawings
Fig. 1 is system structure diagram；
Fig. 2 is method flow diagram；
Fig. 3 Fig. 4 is the diagram signal in method practical implementation.
Embodiment
Below in conjunction with the accompanying drawings and embodiment the invention will be further described.
As shown in figure 1, the present invention provides a kind of character relation digging system based on face identification system, including video is adopted
Collect module, characteristic extracting module, data memory module, data processing module；
The video acquisition module is the video acquisition device in face identification system, is generally laid in designated area
Camera, in real time gather specified range in video frame images；
The characteristic extracting module, for carrying out feature respectively to the multiple different faces occurred in each video frame images
Extract, the corresponding face characteristic vector of every face of generation simultaneously outputs this to data memory module；For being looked into input
The face for asking people carries out feature extraction, and generation is queried the corresponding face characteristic vector of people；
The data memory module is a face information database, the every record correspondence frame of video figure stored in database
The corresponding face information of each personage occurred as in, described face information includes face characteristic vector, face picture, people
The true picture that face occurs, video-frequency band and the unique frame number of total system that face occurs；
The data processing module, for every record to being queried in the characteristic vector and face information database of people
Characteristic vector enter row distance matching, return to qualified face record；For being gathered to qualified face record
Class processing, generates and records the face information table constituted by multigroup class；For according in every group of class record record object number, it is right
Multigroup class record is ranked up and exported.
Using said system, a kind of character relation method for digging based on face identification system of offer of the invention, including with
Lower step：
S1, obtains the video frame images sequence of video acquisition system collection in real time, multiple to what is occurred in each two field picture
Face picture carries out feature extraction respectively, the corresponding characteristic vector of every face of generation, and by this feature vector and corresponding people
The face information database of face information storage in the server；
S2, input is queried human face photo, and the face part on photo is cut out by detection, corrects and ajusts face,
Then feature extraction is carried out in input deep learning neural network algorithm CNN, obtains being queried the characteristic vector of people；
S3, the corresponding characteristic vector of every face picture on face information database and the characteristic vector being queried are carried out
Distance matching, if returning to face picture correspondence video frame number if less than predetermined threshold；Described distance can be remaining
Revolve distance, Euclidean distance, correlation etc..
S4, is inquired about in face information database, obtains and remove with being looked into successively according to the step S3 frame numbers returned
The distance of characteristic vector of people is ask less than other all frame number identical faces records beyond the face record of predetermined threshold value,
And clustering processing is carried out, the class record sheet being made up of the corresponding class record of different target personage is generated, while calculating every group of class note
The central feature vector of record；
S5, is recorded for every group of class, every center for recording corresponding characteristic vector and this group of class record is calculated respectively special
The distance of vector is levied, the record that distance is more than predetermined threshold value is deleted；
S6, record according to included in every group of class record number according to descending sort and export；
S7, when the freshly harvested face picture information of face information database acquisition video acquisition system, by what is be queried
Characteristic vector characteristic vector corresponding with freshly harvested face picture enters row distance matching, if returning if less than predetermined threshold
The corresponding frame number of the face, and extract the corresponding characteristic vector of face picture of all identical frame numbers；
S8, for each characteristic vector, ergodic classes record sheet, if this feature vector and the group class record of certain in class record sheet
The distance of central feature vector be less than predetermined threshold value, then this feature vector corresponding face record is included into this group of class record
And recalculate the central feature vector of this group of class record；Otherwise it is the corresponding face record of this feature vector in class record sheet
A new class is created, while initializing the central feature vector of new class using this feature vector.
The part not illustrated in specification is prior art or common knowledge.The present embodiment is merely to illustrate the invention,
Rather than limitation the scope of the present invention, those skilled in the art change for equivalent replacement of the invention made etc. to be considered
Fall into invention claims institute protection domain.
Claims (7)
1. a kind of character relation digging system based on face identification system, it is characterised in that：Including video acquisition module, feature
Extraction module, data memory module, data processing module；
The video acquisition module is the video acquisition device in face identification system, for gathering regarding in specified range in real time
Frequency two field picture；
The characteristic extracting module, is carried for carrying out feature respectively to the multiple different faces occurred in each video frame images
Take, the corresponding face characteristic vector of every face of generation simultaneously outputs this to data memory module；For being queried to input
The face of people carries out feature extraction, and generation is queried the corresponding face characteristic vector of people；
The data memory module is a face information database, in the every record correspondence video frame images stored in database
The corresponding face information of each personage occurred, described face information includes face characteristic vector, and face picture, face goes out
Existing true picture, video-frequency band and the unique frame number of total system that face occurs；
The data processing module, the spy for every record to being queried in the characteristic vector and face information database of people
Levy vector and enter row distance matching, return to qualified face record；For being carried out to qualified face record at cluster
Reason, generates and records the face information table constituted by multigroup class；For according in every group of class record record object number, to multigroup
Class record is ranked up and exported.
2. a kind of character relation method for digging based on face identification system, it is characterised in that：Comprise the following steps：
S1, obtains the video frame images sequence of video acquisition system collection, to the multiple faces occurred in each two field picture in real time
Picture carries out feature extraction, the corresponding characteristic vector of every face of generation respectively, and this feature vector and corresponding face are believed
The face information database of breath storage in the server；
S2, input is queried human face photo, and the face part on photo is cut out by detection, corrects and ajusts face, then
Feature extraction is carried out, obtains being queried the characteristic vector of people；
S3, row distance is entered by the corresponding characteristic vector of every face picture on face information database and the characteristic vector being queried
Matching, if returning to face picture correspondence video frame number if less than predetermined threshold；
S4, is inquired about in face information database, obtains and remove with being queried people successively according to the step S3 frame numbers returned
The distance of characteristic vector be less than other all frame number identical faces records beyond the face of predetermined threshold value record, go forward side by side
Row clustering processing, generates the class record sheet being made up of the corresponding class record of different target personage, while calculating what every group of class was recorded
Central feature vector；
S5, is recorded for every group of class, and every central feature for recording corresponding characteristic vector and this group of class record is calculated respectively and is sweared
The distance of amount, deletes the record that distance is more than predetermined threshold value；
S6, record according to included in every group of class record number according to descending sort and export.
3. a kind of character relation method for digging based on face identification system according to claim 2, it is characterised in that：Also
Including
S7, when the freshly harvested face picture information of face information database acquisition video acquisition system, by the feature being queried
Vector characteristic vector corresponding with freshly harvested face picture enters row distance matching, if returning to the people if less than predetermined threshold
The corresponding frame number of face, and extract the corresponding characteristic vector of face picture of all identical frame numbers；
S8, for each characteristic vector, ergodic classes record sheet, if during this feature vector is recorded with the group class of certain in class record sheet
The distance of heart characteristic vector is less than predetermined threshold value, then the corresponding face record of this feature vector is included into this group of class record；It is no
A new class then is created for the corresponding face record of this feature vector in class record sheet, while being initialized using this feature vector
The central feature vector of new class.
4. a kind of method that whole face is reduced from local facial region according to claim 3, it is characterised in that：It is described
Step S8 also includes：Preset when the distance of the central feature vector of the group class record of certain in new characteristic vector and class record sheet is less than
Threshold value, and the corresponding face record of this feature vector is included into after this group of class record, the center for recalculating this group of class record is special
Levy vector.
5. a kind of character relation method for digging based on face identification system according to claim any one of 2-4, it is special
Levy and be：The central feature vector is the geometric average of all face characteristic vectors in this group of class record.
6. a kind of character relation method for digging based on face identification system according to claim 5, it is characterised in that：Institute
Feature extraction is stated to realize using deep learning neural network algorithm CNN.
7. a kind of character relation method for digging based on face identification system according to claim 6, it is characterised in that：People
The face information of face information data library storage at least includes herein below：Face picture, the frame picture that face occurs, face occurs
Video-frequency band and the unique number of frames of total system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710220835.1A CN106980844A (en) | 2017-04-06 | 2017-04-06 | A kind of character relation digging system and method based on face identification system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710220835.1A CN106980844A (en) | 2017-04-06 | 2017-04-06 | A kind of character relation digging system and method based on face identification system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106980844A true CN106980844A (en) | 2017-07-25 |
Family
ID=59343744
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710220835.1A Pending CN106980844A (en) | 2017-04-06 | 2017-04-06 | A kind of character relation digging system and method based on face identification system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106980844A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107704851A (en) * | 2017-10-30 | 2018-02-16 | 歌尔股份有限公司 | Character recognition method, Public Media exhibiting device, server and system |
| CN108229314A (en) * | 2017-11-28 | 2018-06-29 | 深圳市商汤科技有限公司 | Searching method, device and the electronic equipment of target person |
| CN109117714A (en) * | 2018-06-27 | 2019-01-01 | 北京旷视科技有限公司 | A kind of colleague's personal identification method, apparatus, system and computer storage medium |
| CN110443906A (en) * | 2019-07-25 | 2019-11-12 | 中新智擎科技有限公司 | A kind of method, device and mobile terminal that attendance information is presented |
| CN110688510A (en) * | 2018-06-20 | 2020-01-14 | 浙江宇视科技有限公司 | Face background image acquisition method and system |
| CN110751043A (en) * | 2019-09-19 | 2020-02-04 | 平安科技（深圳）有限公司 | Face recognition method and device based on face visibility and storage medium |
| CN111627125A (en) * | 2020-06-02 | 2020-09-04 | 上海商汤智能科技有限公司 | Sign-in method, device, computer equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102096800A (en) * | 2009-12-14 | 2011-06-15 | 北京中星微电子有限公司 | Method and device for acquiring image information |
| CN103679151A (en) * | 2013-12-19 | 2014-03-26 | 成都品果科技有限公司 | LBP and Gabor characteristic fused face clustering method |
| CN104133875A (en) * | 2014-07-24 | 2014-11-05 | 北京中视广信科技有限公司 | Face-based video labeling method and face-based video retrieving method |
| CN105740842A (en) * | 2016-03-01 | 2016-07-06 | 浙江工业大学 | Unsupervised face recognition method based on fast density clustering algorithm |
| CN106228120A (en) * | 2016-07-14 | 2016-12-14 | 南京航空航天大学 | The extensive human face data mask method of query driven |
-
2017
- 2017-04-06 CN CN201710220835.1A patent/CN106980844A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102096800A (en) * | 2009-12-14 | 2011-06-15 | 北京中星微电子有限公司 | Method and device for acquiring image information |
| CN103679151A (en) * | 2013-12-19 | 2014-03-26 | 成都品果科技有限公司 | LBP and Gabor characteristic fused face clustering method |
| CN104133875A (en) * | 2014-07-24 | 2014-11-05 | 北京中视广信科技有限公司 | Face-based video labeling method and face-based video retrieving method |
| CN105740842A (en) * | 2016-03-01 | 2016-07-06 | 浙江工业大学 | Unsupervised face recognition method based on fast density clustering algorithm |
| CN106228120A (en) * | 2016-07-14 | 2016-12-14 | 南京航空航天大学 | The extensive human face data mask method of query driven |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107704851A (en) * | 2017-10-30 | 2018-02-16 | 歌尔股份有限公司 | Character recognition method, Public Media exhibiting device, server and system |
| CN107704851B (en) * | 2017-10-30 | 2021-01-15 | 歌尔股份有限公司 | Character identification method, public media display device, server and system |
| US10891465B2 (en) | 2017-11-28 | 2021-01-12 | Shenzhen Sensetime Technology Co., Ltd. | Methods and apparatuses for searching for target person, devices, and media |
| CN108229314A (en) * | 2017-11-28 | 2018-06-29 | 深圳市商汤科技有限公司 | Searching method, device and the electronic equipment of target person |
| CN110431560B (en) * | 2017-11-28 | 2023-12-19 | 深圳市商汤科技有限公司 | Target person searching method, device, equipment and medium |
| WO2019105163A1 (en) * | 2017-11-28 | 2019-06-06 | 深圳市商汤科技有限公司 | Target person search method and apparatus, device, program product and medium |
| CN110431560A (en) * | 2017-11-28 | 2019-11-08 | 深圳市商汤科技有限公司 | The searching method and device of target person, equipment, program product and medium |
| CN110688510B (en) * | 2018-06-20 | 2022-06-14 | 浙江宇视科技有限公司 | Face background image acquisition method and system |
| CN110688510A (en) * | 2018-06-20 | 2020-01-14 | 浙江宇视科技有限公司 | Face background image acquisition method and system |
| CN109117714A (en) * | 2018-06-27 | 2019-01-01 | 北京旷视科技有限公司 | A kind of colleague's personal identification method, apparatus, system and computer storage medium |
| CN110443906A (en) * | 2019-07-25 | 2019-11-12 | 中新智擎科技有限公司 | A kind of method, device and mobile terminal that attendance information is presented |
| CN110751043A (en) * | 2019-09-19 | 2020-02-04 | 平安科技（深圳）有限公司 | Face recognition method and device based on face visibility and storage medium |
| WO2021051611A1 (en) * | 2019-09-19 | 2021-03-25 | 平安科技（深圳）有限公司 | Face visibility-based face recognition method, system, device, and storage medium |
| CN110751043B (en) * | 2019-09-19 | 2023-08-22 | 平安科技（深圳）有限公司 | Face recognition method and device based on face visibility and storage medium |
| CN111627125A (en) * | 2020-06-02 | 2020-09-04 | 上海商汤智能科技有限公司 | Sign-in method, device, computer equipment and storage medium |
| CN111627125B (en) * | 2020-06-02 | 2022-09-27 | 上海商汤智能科技有限公司 | Sign-in method, device, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106980844A (en) | A kind of character relation digging system and method based on face identification system | |
| Kong et al. | Mmact: A large-scale dataset for cross modal human action understanding | |
| CN109522853B (en) | Face datection and searching method towards monitor video | |
| Hamdoun et al. | Person re-identification in multi-camera system by signature based on interest point descriptors collected on short video sequences | |
| Jiang et al. | Recognizing human actions by learning and matching shape-motion prototype trees | |
| US9171197B2 (en) | Facial tracking method | |
| US7505621B1 (en) | Demographic classification using image components | |
| Wang et al. | P2snet: Can an image match a video for person re-identification in an end-to-end way? | |
| CN110263613A (en) | Monitor video processing method and processing device | |
| Asif et al. | Privacy preserving human fall detection using video data | |
| CN105184238A (en) | Human face recognition method and system | |
| CN107230267A (en) | Intelligence In Baogang Kindergarten based on face recognition algorithms is registered method | |
| CN102682302A (en) | Human body posture identification method based on multi-characteristic fusion of key frame | |
| CN107315795B (en) | The instance of video search method and system of joint particular persons and scene | |
| CN103996046A (en) | Personnel recognition method based on multi-visual-feature fusion | |
| CN106845513B (en) | Manpower detector and method based on condition random forest | |
| CN110348347A (en) | A kind of information processing method and device, storage medium | |
| CN107392182A (en) | A kind of face collection and recognition method and device based on deep learning | |
| CN107292228A (en) | A kind of method for accelerating face recognition search speed | |
| Sapkota et al. | Large scale unconstrained open set face database | |
| Liu et al. | Dual context-aware refinement network for person search | |
| KR101170676B1 (en) | Face searching system and method based on face recognition | |
| Wang et al. | Beyond pedestrians: A hybrid approach of tracking multiple articulating humans | |
| CN106557523A (en) | Presentation graphics system of selection and equipment and object images search method and equipment | |
| CN110287841B (en) | Image transmission method and apparatus, image transmission system, and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170725 |
|
| RJ01 | Rejection of invention patent application after publication |