CN102609695A - Method and system for recognizing human face from multiple angles - Google Patents
Method and system for recognizing human face from multiple angles Download PDFInfo
- Publication number
- CN102609695A CN102609695A CN2012100328375A CN201210032837A CN102609695A CN 102609695 A CN102609695 A CN 102609695A CN 2012100328375 A CN2012100328375 A CN 2012100328375A CN 201210032837 A CN201210032837 A CN 201210032837A CN 102609695 A CN102609695 A CN 102609695A
- Authority
- CN
- China
- Prior art keywords
- face
- people
- picture
- face picture
- image information
- 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
Images
Landscapes
- Image Processing (AREA)
Abstract
The invention provides a method and system of recognizing the human face from multiple angles. The method comprises the following steps: acquiring image information from multiple angles, detecting the image information, collecting human face images from the image information, receiving the human face images, distinguishing the attitudes of the human faces, selecting front human face images, extracting the feature values of the front human face images in the collection period, classifying the human faces, and outputting. The human face images of the collected object can be collected under the condition that the collected object does not know and does not need to cooperate, and the front human face image collection rate can be improved by collecting from multiple angels under the condition that the attitudes of the human faces change randomly.
Description
Technical field
The present invention relates to the computer image processing technology field, particularly a kind of multi-angle face identification method and system.
Background technology
Now, face recognition technology is applied in the video safety defense monitoring system gradually, and it uses prerequisite is the collection of people's face.Collecting front face is the application foundation of face recognition technology, but in reality, people's attitude has very big randomness; Can also force by the attitude of acquisition target in occasions such as safety check, entry and exit, but in the public cultural facilities place, for example museum, the Art Museum, library etc.; Can't force spectators to cooperate collection; And in the practical application of video safety defense monitoring system, implement hidden collection often, not perception of acquisition target, so the attitude of acquisition target is uncontrollable.Present people's face acquisition method often is difficult to collect front face in this case, and it is very high to show that mainly present face recognition technology requires the collection front face, and the face rotating range of asking for help is little; Often acquisition camera visual field central axis about [15 °; + 15 °], left rotation and right rotation [20 ° ,+20 °]; And actual persons face or human body attitude scope all exceed this scope, in case exceed then can't collect front face.Owing to people's attitude is random, the installation site of acquisition camera often is difficult to accurately location, and aligning the collection of dough figurine face influences bigger.Present acquisition method often requires to be met the requirement of front face acquisition range by the acquisition target human face posture, needs acquisition target to cooperate, and acquisition target cooperates and hidden collection situation front face acquisition rate is not high for need not.
The multiple attitude recognition of face that discloses like the Chinese patent CN1004565 one predetermined threshold value system and method for deploying to ensure effective monitoring and control of illegal activities; Though its can know others face rotating range be about [30 °; + 30 °], left rotation and right rotation [30 ° ,+30 °]; In case but the rotation of people's face exceeds above scope, still can't realize effective collection.And under the ill-matched situation of not perception of acquisition target, human face posture often all can exceed above scope.
To deficiency of the prior art; Propose a kind of ability and realize people's face collection being need not to cooperate by the acquisition target unaware; To improve the multi-angle face identification method and the system of front face acquisition rate under the bigger situation of human face posture randomness, is one of computer image processing technology field problem anxious to be solved at present through the multi-angle collection.
Summary of the invention
In view of this; The embodiment of the invention has proposed a kind of multi-angle face identification method and system, obtains image information through a plurality of angles, then image information is detected and gathers people's face picture wherein; And through recipient's face picture and it is carried out attitude differentiate; Select the front face picture, finally the front face picture in the collection period is extracted eigenwert, carry out people's face clustering processing and output; The people face collection of realization to being need not to cooperate by the acquisition target unaware, through the multi-angle collection to improve front face acquisition rate under the bigger situation of human face posture randomness.
For solving the problems of the technologies described above, the purpose of the embodiment of the invention realizes through following technical scheme:
A kind of multi-angle face identification method comprises:
Step 1, obtain image information through a plurality of angles;
Step 2, image information is detected and gathers people's face picture wherein;
Step 3, recipient's face picture also carry out attitude to it and differentiate, and select the front face picture;
Step 4, the front face picture in the collection period is extracted eigenwert, carry out people's face clustering processing and output.
Preferably, in the above-mentioned steps one, be the image information of obtaining a plurality of angles through many digital collection video cameras.
Preferably, the selection of above-mentioned digital collection video camera seat in the plane needs according to current face recognition algorithms characteristics, and promptly people from system face can be discerned the attitude scope: left rotation and right rotation [20 ° ,+20 °]; Last lower angle [15 ° ,+15 °], setting height(from bottom) is controlled in the 2700mm; Vertical angle is controlled in the luffing angle [15 ° ,+15 °], axis is installed along personnel's direct of travel; Horizontal sextant angle is no more than 20 degree, is controlled within human face posture left rotation and right rotation [20 °, the 20 °] scope.
Preferably, can gather left rotation and right rotation [60 ° ,+60 °] through 3 video cameras, rotate people's face within [25 ° ,+25 °] scope up and down, the collecting efficiency of every digital camera was 25 frame/seconds.
Preferably, in the above-mentioned steps two, be the video flowing that obtains pictorial information; And video flowing is carried out the frame picture reduce; Restore every video camera per second 25 frame pictures, in the frame picture, carry out people's face subsequently and detect, detect the people of point range intercepting according to pixels face picture and name behind people's face.
Preferably; In the above-mentioned steps two; Be video flowing to be carried out frame picture reduction detect, be pixel coverage intercepting people face picture to people's face of reaching the 100*120 pixel by the 150*250 of people's face center position, and name by the frame picture sequence numbers of image acquisition unit sequence number, collection period time, each collection period reduction acquisition with people's face; Picture size is about 50k, and per second can be gathered people's face picture of the about 3750k of output altogether.
Preferably, in the above-mentioned steps three, be that the people's face picture that receives is differentiated carrying out attitude from each image acquisition unit people's face picture of moment order at same frame together by the image file naming order, select the front face picture.
Preferably, in the above-mentioned steps four, be to extract the face characteristic value, and be that unit compares the face characteristic value in twos that comparison efficiency is about per second 100,000 times by collection period through algorithm.
Preferably, be that lineup's face picture that ratio reaches a predetermined threshold value is judged to be same people through similarity comparison, and choose in immediate two the people's face pictures of similarity ratio one and export as people's face cluster for the face characteristic value.
Preferably, the span of said predetermined threshold value is 60% to 100%.
A kind of multi-angle face identification system; Comprise image acquisition unit, people's face collecting unit, attitude judgement unit and people's face cluster and output unit, obtain image information, then image information is detected and gathers people's face picture wherein through a plurality of angles; And through recipient's face picture and it is carried out attitude differentiate; Select the front face picture, finally the front face picture in the collection period is extracted eigenwert, carry out people's face clustering processing and output.
Preferably, above-mentioned image acquisition unit is used for obtaining image information through a plurality of angles.
Preferably, above-mentioned people's face collecting unit is used for image information is detected and gather people's face picture wherein.
Preferably, above-mentioned attitude judgement unit is used for recipient's face picture and it is carried out attitude differentiate, and selects the front face picture.
Preferably, above-mentioned people's face cluster and output unit are used for the front face picture in the collection period is extracted eigenwert, carry out people's face clustering processing and output.
In sum; The invention provides a kind of multi-angle face identification method and system, obtain image information, then image information is detected and gathers people's face picture wherein through a plurality of angles; And through recipient's face picture and it is carried out attitude differentiate; Select the front face picture, finally the front face picture in the collection period is extracted eigenwert, carry out people's face clustering processing and output; The people face collection of realization to being need not to cooperate by the acquisition target unaware, through the multi-angle collection to improve front face acquisition rate under the bigger situation of human face posture randomness.
Description of drawings
Fig. 1 is a kind of multi-angle face identification method process flow diagram of the embodiment of the invention;
Fig. 2 installs vertical view for angle video camera more than the present invention's one specific embodiment;
Fig. 3 is the video camera installation side view of invention one specific embodiment;
A kind of multi-angle face identification system synoptic diagram that Fig. 4 provides for the embodiment of the invention.
Embodiment
A kind of multi-angle face identification method and system that the embodiment of the invention provides; Obtain image information through a plurality of angles; Then image information is detected and gathers people's face picture wherein, and carry out the attitude differentiation, select the front face picture through recipient's face picture and to it; Finally the front face picture in the collection period is extracted eigenwert; Carry out people's face clustering processing and output, realize people's face collection being need not to cooperate by the acquisition target unaware, through the multi-angle collection to improve front face acquisition rate under the bigger situation of human face posture randomness.
For making the object of the invention, technical scheme and advantage clearer, the embodiment that develops simultaneously with reference to the accompanying drawings is to further explain of the present invention.
The embodiment of the invention provides a kind of multi-angle face identification method, and as shown in Figure 1, concrete steps comprise:
Step 1, obtain image information through a plurality of angles;
Particularly, in embodiments of the present invention, comprising a plurality of image acquisition units so that obtain image information from a plurality of angles, is the image information of obtaining a plurality of angles through many digital collection video cameras.In this programme, comprise image acquisition unit one, image acquisition unit two, image acquisition unit three and be the digital collection video camera, be used for multi-angle to obtain image its scheme of installation such as Fig. 2, shown in Figure 3.The selection of every video camera seat in the plane needs according to current face recognition algorithms characteristics, and promptly people from system face can be discerned the attitude scope: left rotation and right rotation [20 ° ,+20 °], last lower angle [15 ° ,+15 °].Setting height(from bottom) is controlled in the 2700mm, and vertical angle is controlled in the luffing angle [15 ° ,+15 °], axis is installed along personnel's direct of travel, and horizontal sextant angle is no more than 20 degree, is controlled within human face posture left rotation and right rotation [20 °, the 20 °] scope.Can gather left rotation and right rotation [60 ° ,+60 °] through 3 video cameras, up and down people's face within rotation [25 ° ,+25 °] scope.The collecting efficiency of every digital camera was 25 frame/seconds.
Further, in this programme, image acquisition unit one is a digital camera, and for people's face collecting unit provides image, its setting height(from bottom) is as shown in Figure 3, and collecting efficiency was 25 frame/seconds.Image acquisition unit two is a digital camera, and for people's face collecting unit provides image, its setting height(from bottom) is as shown in Figure 3, and collecting efficiency was 25 frame/seconds.Image acquisition unit three is a digital camera, and for people's face collecting unit provides image, its setting height(from bottom) is as shown in Figure 3, and collecting efficiency was 25 frame/seconds.Image acquisition unit is installed vertical view such as Fig. 2.
Step 2, image information is detected and gathers people's face picture wherein;
Particularly, in embodiments of the present invention, will detect and gather people's face picture wherein to the image information of obtaining through a plurality of angles.In this programme; At first will obtain the video flowing of each image acquisition unit output; Secondly video flowing is carried out the reduction of frame picture, restore every video camera per second 25 frame pictures, in the frame picture, carry out people's face subsequently and detect; The people of point range intercepting according to pixels face picture is also named after detecting people's face, and its method is following:
Collection period of define system is 1 second, gathers first to be T0, and next collection period is T1, by that analogy.
Definition image acquisition unit N people's face picture to be detected and intercepting in Tn moment P two field picture is R
NnP, the people's face picture that then is detected constantly at T0 is following:
For image acquisition unit one: R
101R
102R
103... R
1025
For image acquisition unit two: R
201R
202R
203... R
2025
For image acquisition unit three: R
301R
302R
303... R
3025
Further; In this programme, the people's face collecting unit that comprises at first obtains the video flowing of each image acquisition unit output, secondly video flowing is carried out the reduction of frame picture; Restore every video camera per second 25 frame pictures; In the frame picture, carry out people's face subsequently and detect, detect the people of point range intercepting according to pixels face picture and name behind people's face, and export to the attitude judgement unit.People's face collecting unit is a DSP microchip processor; Can obtain the video flowing of image acquisition unit in real time; Video flowing is carried out reduction of frame picture and the detection of people's face; Is pixel coverage intercepting people face picture to people's face of reaching the 100*120 pixel by the 150*250 of people's face center position, and names by the frame picture sequence numbers that image acquisition unit sequence number, collection period time, each collection period reduction obtain, and picture size is about 50k.Per second can be gathered people's face picture of the about 3750k of output altogether.
Step 3, recipient's face picture also carry out attitude to it and differentiate, and select the front face picture;
Particularly, in embodiments of the present invention, because people's face can only have one to be that unit carries out algorithm and differentiates towards, the people's face picture that therefore a certain moment is obtained by each image acquisition unit frame picture in a certain particular moment, and at R
101, R
201, R
301In determine a front face, by that analogy.
Further; It is DSP microchip processor that this programme comprises an attitude judgement unit; With people's face picture output of recipient's face collecting unit, differentiate carrying out attitude by the image file naming order from each image acquisition unit people's face picture of moment order at same frame together, select the front face picture.Attitude differentiation algorithm operational efficiency is about per second and differentiates 10,000 people's face pictures.Front face through identification and classification goes out is exported to people's face cluster cell.
Step 4, the front face picture in the collection period is extracted eigenwert, carry out people's face clustering processing and output.
Particularly, in embodiments of the present invention, will extract eigenwert to one group of front face of a certain collection period human face posture judgement unit output, and carry out the similarity comparison, the cluster that similarity ratio reaches a predetermined threshold value is people's face.In this programme; The front face that comprises a pair of T0 collection period of people's face cluster cell carries out cluster; The front face of two pairs of T1 collection period of people's face cluster cell carries out cluster; People's face cluster cell carries out cluster to the front face of T2 collection period again and again, and people's face cluster cell two carries out cluster to the front face of T3 collection period again, by that analogy.Because the time that people's face occurs before acquisition camera in the reality was greater than 1 second collection period; And, through being people's face of many different people after the cluster, therefore; The front face that people's face cluster cell one cluster is accomplished also need carry out the secondary cluster with the front face that next collection period cluster is accomplished; Promptly give people's face cluster cell two cluster once more, reciprocal successively, the same front face of each collection period can be come to light; And to front face output of the final formation of different objects, the span of said predetermined threshold value is 60% to 100%.。
Further, in this programme, the people's face cluster cell that comprises is an application server, extracts the face characteristic value through algorithm, and is that unit compares the face characteristic value in twos by collection period, and comparison efficiency is about per second 100,000 times.Through similarity comparison, lineup's face picture that ratio reaches a predetermined threshold value is judged to be same people, and choose in immediate two the people's face pictures of similarity ratio one and export as people's face cluster for the face characteristic value.The output of people's face cluster cell one is as an importation of people's face cluster cell two; Participate in people's face cluster of next collection period; People's face cluster cell one and back and forth exchange input and output of people's face cluster cell two; No longer close with other people face eigenwert until this face characteristic value, promptly ratio is exported this people's face picture and eigenwert less than 0.6.Can concentrate cluster to M collection period according to actual conditions setup parameter M during practical application, running efficiency of system is provided.In this programme, the front face output form of output unit is picture and this picture characteristic of correspondence value coding.
In addition, the embodiment of the invention also provides a kind of multi-angle face identification system.A kind of multi-angle face identification system synoptic diagram as shown in Figure 4, as to provide for the embodiment of the invention.
A kind of multi-angle face identification system comprises image acquisition unit 11, people's face collecting unit 22, attitude judgement unit 33 and people's face cluster and output unit 44.
Particularly, in embodiments of the present invention, comprise a plurality of image acquisition units so that obtain image information from a plurality of angles.In this programme, comprise image acquisition unit one, image acquisition unit two, image acquisition unit three and be the digital collection video camera, be used for multi-angle to obtain image its scheme of installation such as Fig. 2, shown in Figure 3.The selection of every video camera seat in the plane needs according to current face recognition algorithms characteristics, and promptly people from system face can be discerned the attitude scope: left rotation and right rotation [20 ° ,+20 °], last lower angle [15 ° ,+15 °].Setting height(from bottom) is controlled in the 2700mm, and vertical angle is controlled in the luffing angle [15 ° ,+15 °], axis is installed along personnel's direct of travel, and horizontal sextant angle is no more than 20 degree, is controlled within human face posture left rotation and right rotation [20 °, the 20 °] scope.Can gather left rotation and right rotation [60 ° ,+60 °] through 3 video cameras, up and down people's face within rotation [25 ° ,+25 °] scope.The collecting efficiency of every digital camera was 25 frame/seconds.
Further, in this programme, image acquisition unit one is a digital camera, and for people's face collecting unit provides image, its setting height(from bottom) is as shown in Figure 3, and collecting efficiency was 25 frame/seconds.Image acquisition unit two is a digital camera, and for people's face collecting unit provides image, its setting height(from bottom) is as shown in Figure 3, and collecting efficiency was 25 frame/seconds.Image acquisition unit three is a digital camera, and for people's face collecting unit provides image, its setting height(from bottom) is as shown in Figure 3, and collecting efficiency was 25 frame/seconds.Image acquisition unit is installed vertical view such as Fig. 2.
People's face collecting unit 22 is used for image information is detected and gather people's face picture wherein;
Particularly, in embodiments of the present invention, will detect and gather people's face picture wherein to the image information of obtaining through a plurality of angles.In this programme; At first will obtain the video flowing of each image acquisition unit output; Secondly video flowing is carried out the reduction of frame picture, restore every video camera per second 25 frame pictures, in the frame picture, carry out people's face subsequently and detect; The people of point range intercepting according to pixels face picture is also named after detecting people's face, and its method is following:
Collection period of define system is 1 second, gathers first to be T0, and next collection period is T1, by that analogy.
Definition image acquisition unit N people's face picture to be detected and intercepting in Tn moment P two field picture is R
NnP, the people's face picture that then is detected constantly at T0 is following:
For image acquisition unit one: R
101R
102R
103... R
1025
For image acquisition unit two: R
201R
202R
203... R
2025
For image acquisition unit three: R
301R
302R
303... R
3025
Further; In this programme, the people's face collecting unit that comprises at first obtains the video flowing of each image acquisition unit output, secondly video flowing is carried out the reduction of frame picture; Restore every video camera per second 25 frame pictures; In the frame picture, carry out people's face subsequently and detect, detect the people of point range intercepting according to pixels face picture and name behind people's face, and export to the attitude judgement unit.People's face collecting unit is a DSP microchip processor; Can obtain the video flowing of image acquisition unit in real time; Video flowing is carried out reduction of frame picture and the detection of people's face; Is pixel coverage intercepting people face picture to people's face of reaching the 100*120 pixel by the 150*250 of people's face center position, and names by the frame picture sequence numbers that image acquisition unit sequence number, collection period time, each collection period reduction obtain, and picture size is about 50k.Per second can be gathered people's face picture of the about 3750k of output altogether.
Particularly, in embodiments of the present invention, because people's face can only have one to be that unit carries out algorithm and differentiates towards, the people's face picture that therefore a certain moment is obtained by each image acquisition unit frame picture in a certain particular moment, and at R
101, R
201, R
301In determine a front face, by that analogy.
Further; It is DSP microchip processor that this programme comprises an attitude judgement unit; With people's face picture output of recipient's face collecting unit, differentiate carrying out attitude by the image file naming order from each image acquisition unit people's face picture of moment order at same frame together, select the front face picture.Attitude differentiation algorithm operational efficiency is about per second and differentiates 10,000 people's face pictures.Front face through identification and classification goes out is exported to people's face cluster cell.
People's face cluster and output unit 44 are used for the front face picture in the collection period is extracted eigenwert, carry out people's face clustering processing and output.
Particularly, in embodiments of the present invention, will extract eigenwert to one group of front face of a certain collection period human face posture judgement unit output, and carry out the similarity comparison, the cluster that similarity ratio reaches a predetermined threshold value is people's face.In this programme; The front face that comprises a pair of T0 collection period of people's face cluster cell carries out cluster; The front face of two pairs of T1 collection period of people's face cluster cell carries out cluster; People's face cluster cell carries out cluster to the front face of T2 collection period again and again, and people's face cluster cell two carries out cluster to the front face of T3 collection period again, by that analogy.Because the time that people's face occurs before acquisition camera in the reality was greater than 1 second collection period; And, through being people's face of many different people after the cluster, therefore; The front face that people's face cluster cell one cluster is accomplished also need carry out the secondary cluster with the front face that next collection period cluster is accomplished; Promptly give people's face cluster cell two cluster once more, reciprocal successively, the same front face of each collection period can be come to light; And to front face output of the final formation of different objects, the span of said predetermined threshold value is 60% to 100%.。
Further, in this programme, the people's face cluster cell that comprises is an application server, extracts the face characteristic value through algorithm, and is that unit compares the face characteristic value in twos by collection period, and comparison efficiency is about per second 100,000 times.Through similarity comparison, lineup's face picture that ratio reaches a predetermined threshold value is judged to be same people, and choose in immediate two the people's face pictures of similarity ratio one and export as people's face cluster for the face characteristic value.The output of people's face cluster cell one is as an importation of people's face cluster cell two; Participate in people's face cluster of next collection period; People's face cluster cell one and back and forth exchange input and output of people's face cluster cell two; No longer close with other people face eigenwert until this face characteristic value, promptly ratio is exported this people's face picture and eigenwert less than 0.6.Can concentrate cluster to M collection period according to actual conditions setup parameter M during practical application, running efficiency of system is provided.In this programme, the front face output form of output unit is picture and this picture characteristic of correspondence value coding.
Through the multi-angle face identification method and the system of this programme, can realize people's face collection to bigger human face posture scope, remedy the deficiency of current face recognition algorithms in practical application; Make people's face can gather the attitude scope and extend to [60 ° ,+60 °] by left rotation and right rotation [20 ° ,+20 °]; Rotating range is by [15 ° up and down; + 15 °] extend to [25 ° ,+25 °], improve the front face collecting efficiency.On the other hand, realized being need not collection of people's face and cluster under the mated condition by the acquisition target unaware.
One of ordinary skill in the art will appreciate that and realize that all or part of step that the foregoing description method is carried is to instruct relevant hardware to accomplish through program; Described program can be stored in a kind of computer-readable recording medium; This program comprises one of step or its combination of method embodiment when carrying out.
In addition, each functional unit in each embodiment of the present invention can be integrated in the processing module, also can be that the independent physics in each unit exists, and also can be integrated in the module two or more unit.Above-mentioned integrated module both can adopt the form of hardware to realize, also can adopt the form of software function module to realize.If said integrated module realizes with the form of software function module and during as independently production marketing or use, also can be stored in the computer read/write memory medium.
In sum; This paper provides a kind of multi-angle face identification method and system, obtains image information through a plurality of angles, then image information is detected and gathers people's face picture wherein; And through recipient's face picture and it is carried out attitude differentiate; Select the front face picture, finally the front face picture in the collection period is extracted eigenwert, carry out people's face clustering processing and output; The people face collection of realization to being need not to cooperate by the acquisition target unaware, through the multi-angle collection to improve front face acquisition rate under the bigger situation of human face posture randomness.
More than a kind of multi-angle face identification method provided by the present invention and system have been carried out detailed introduction; Used concrete example among this paper principle of the present invention and embodiment are set forth, the explanation of above embodiment just is used for helping to understand scheme of the present invention; Simultaneously, for one of ordinary skill in the art, according to thought of the present invention, the part that on embodiment and range of application, all can change, in sum, this description should not be construed as limitation of the present invention.
Claims (15)
1. a multi-angle face identification method is characterized in that, said method comprises:
Step 1, collude.Cross a plurality of angles and obtain image information;
Step 2, image information is detected and gathers people's face picture wherein;
Step 3, recipient's face picture also carry out attitude to it and differentiate, and select the front face picture;
Step 4, the front face picture in the collection period is extracted eigenwert, carry out people's face clustering processing and output.
2. method according to claim 1 is characterized in that, in the said step 1, is the image information of obtaining a plurality of angles through many digital collection video cameras.
3. method according to claim 2 is characterized in that, the selection of said digital collection video camera seat in the plane needs according to current face recognition algorithms characteristics; Be that people from system face can be discerned the attitude scope: left rotation and right rotation [20 ° ,+20 °], [15 ° of last lower angles; + 15 °], setting height(from bottom) is controlled in the 2700mm, and vertical angle is controlled at [15 ° of luffing angles; + 15 °] in, axis is installed along personnel's direct of travel, horizontal sextant angle is no more than 20 degree; Be controlled within human face posture left rotation and right rotation [20 °, the 20 °] scope.
4. method according to claim 2 is characterized in that, can gather left rotation and right rotation [60 ° ,+60 °] through 3 video cameras, rotates people's face within [25 ° ,+25 °] scope up and down, and the collecting efficiency of every digital camera was 25 frame/seconds.
5. method according to claim 1 and 2; It is characterized in that, in the said step 2, be the video flowing that obtains pictorial information; And video flowing is carried out the frame picture reduce; Restore every video camera per second 25 frame pictures, in the frame picture, carry out people's face subsequently and detect, detect the people of point range intercepting according to pixels face picture and name behind people's face.
6. method according to claim 1; It is characterized in that, in the said step 2, be that video flowing is carried out reduction of frame picture and the detection of people's face; Is pixel coverage intercepting people face picture to people's face of reaching the 100*120 pixel by the 150*250 of people's face center position; And name by the frame picture sequence numbers that image acquisition unit sequence number, collection period time, each collection period reduction obtain, picture size is about 50k, and per second can be gathered people's face picture of the about 3750k of output altogether.
7. method according to claim 1; It is characterized in that; In the said step 3, be that the people's face picture that receives is differentiated carrying out attitude from each image acquisition unit people's face picture of moment order at same frame together by the image file naming order, select the front face picture.
8. method according to claim 1 is characterized in that, in the said step 4, is to extract the face characteristic value through algorithm, and is that unit compares the face characteristic value in twos by collection period, and comparison efficiency is about per second 100,000 times.
9. method according to claim 8; It is characterized in that; Through similarity comparison, lineup's face picture that ratio reaches a predetermined threshold value is judged to be same people, and choose in immediate two the people's face pictures of similarity ratio one and export as people's face cluster for the face characteristic value.
10. method according to claim 9 is characterized in that, the span of said predetermined threshold value is 60% to 100%.
11. a multi-angle face identification system is characterized in that, said system comprises image acquisition unit, people's face collecting unit, attitude judgement unit and people's face cluster and output unit; Obtain image information through a plurality of angles; Then image information is detected and gathers people's face picture wherein, and carry out the attitude differentiation, select the front face picture through recipient's face picture and to it; Finally the front face picture in the collection period is extracted eigenwert, carry out people's face clustering processing and output.
12. system according to claim 11 is characterized in that, said image acquisition unit is used for obtaining image information through a plurality of angles.
13. system according to claim 11 is characterized in that, said people's face collecting unit is used for image information is detected and gather people's face picture wherein.
14. system according to claim 11 is characterized in that, said attitude judgement unit is used for recipient's face picture and it is carried out attitude differentiate, and selects the front face picture.
15. system according to claim 11 is characterized in that, said people's face cluster and output unit are used for the front face picture in the collection period is extracted eigenwert, carry out people's face clustering processing and output.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100328375A CN102609695A (en) | 2012-02-14 | 2012-02-14 | Method and system for recognizing human face from multiple angles |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100328375A CN102609695A (en) | 2012-02-14 | 2012-02-14 | Method and system for recognizing human face from multiple angles |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102609695A true CN102609695A (en) | 2012-07-25 |
Family
ID=46527054
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012100328375A Pending CN102609695A (en) | 2012-02-14 | 2012-02-14 | Method and system for recognizing human face from multiple angles |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102609695A (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103634503A (en) * | 2013-12-16 | 2014-03-12 | 苏州大学 | Video manufacturing method based on face recognition and behavior recognition and video manufacturing method based on face recognition and behavior recognition |
| CN104091149A (en) * | 2014-06-20 | 2014-10-08 | 中国船舶重工集团公司第七二六研究所 | Face acquisition modeling training system and acquisition modeling training method thereof |
| CN105590097A (en) * | 2015-12-17 | 2016-05-18 | 重庆邮电大学 | Security system and method for recognizing face in real time with cooperation of double cameras on dark condition |
| CN106295522A (en) * | 2016-07-29 | 2017-01-04 | 武汉理工大学 | A kind of two-stage anti-fraud detection method based on multi-orientation Face and environmental information |
| CN106503687A (en) * | 2016-11-09 | 2017-03-15 | 合肥工业大学 | The monitor video system for identifying figures of fusion face multi-angle feature and its method |
| CN106657544A (en) * | 2016-10-24 | 2017-05-10 | 广东欧珀移动通信有限公司 | Incoming call recording method and terminal equipment |
| CN107633244A (en) * | 2017-11-13 | 2018-01-26 | 姜茂清 | Human body direct picture harvester, the image capturing system comprising it and its application |
| WO2018040045A1 (en) * | 2016-08-31 | 2018-03-08 | 深圳前海达闼云端智能科技有限公司 | Monitoring method, apparatus and electronic device |
| CN109034178A (en) * | 2018-05-28 | 2018-12-18 | 北京文香信息技术有限公司 | A kind of demographic method based on face characteristic array |
| CN109214324A (en) * | 2018-08-27 | 2019-01-15 | 曜科智能科技（上海）有限公司 | Most face image output method and output system based on polyphaser array |
| CN110338759A (en) * | 2019-06-27 | 2019-10-18 | 嘉兴深拓科技有限公司 | A kind of front pain expression data acquisition method |
| WO2019223313A1 (en) * | 2018-05-22 | 2019-11-28 | 深圳云天励飞技术有限公司 | Personnel file establishment method and apparatus |
| CN110781710A (en) * | 2018-12-17 | 2020-02-11 | 北京嘀嘀无限科技发展有限公司 | Target object clustering method and device |
| CN111444822A (en) * | 2020-03-24 | 2020-07-24 | 北京奇艺世纪科技有限公司 | Object recognition method and apparatus, storage medium, and electronic apparatus |
| CN111539911A (en) * | 2020-03-23 | 2020-08-14 | 中国科学院自动化研究所 | Mouth breathing face recognition method, device and storage medium |
| CN112001360A (en) * | 2020-09-09 | 2020-11-27 | 深圳中神电子科技有限公司 | Face recognition monitoring system based on intelligent adjustment |
| CN113327332A (en) * | 2018-09-29 | 2021-08-31 | 爱笔（北京）智能科技有限公司 | Biological information-based ticket processing method and system, server and client |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040197014A1 (en) * | 2003-04-01 | 2004-10-07 | Honda Motor Co., Ltd. | Face identification system |
| CN101673346A (en) * | 2008-09-09 | 2010-03-17 | 日电（中国）有限公司 | Method, equipment and system for processing image |
| CN201689439U (en) * | 2010-05-20 | 2010-12-29 | 上海洪剑智能科技有限公司 | Distributed face recognizing system |
-
2012
- 2012-02-14 CN CN2012100328375A patent/CN102609695A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040197014A1 (en) * | 2003-04-01 | 2004-10-07 | Honda Motor Co., Ltd. | Face identification system |
| CN101673346A (en) * | 2008-09-09 | 2010-03-17 | 日电（中国）有限公司 | Method, equipment and system for processing image |
| CN201689439U (en) * | 2010-05-20 | 2010-12-29 | 上海洪剑智能科技有限公司 | Distributed face recognizing system |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103634503A (en) * | 2013-12-16 | 2014-03-12 | 苏州大学 | Video manufacturing method based on face recognition and behavior recognition and video manufacturing method based on face recognition and behavior recognition |
| CN104091149A (en) * | 2014-06-20 | 2014-10-08 | 中国船舶重工集团公司第七二六研究所 | Face acquisition modeling training system and acquisition modeling training method thereof |
| CN104091149B (en) * | 2014-06-20 | 2018-01-26 | 中国船舶重工集团公司第七二六研究所 | The collection modeling training system and its collection modeling training method of face |
| CN105590097B (en) * | 2015-12-17 | 2019-01-25 | 重庆邮电大学 | Dual camera collaboration real-time face identification security system and method under the conditions of noctovision |
| CN105590097A (en) * | 2015-12-17 | 2016-05-18 | 重庆邮电大学 | Security system and method for recognizing face in real time with cooperation of double cameras on dark condition |
| CN106295522A (en) * | 2016-07-29 | 2017-01-04 | 武汉理工大学 | A kind of two-stage anti-fraud detection method based on multi-orientation Face and environmental information |
| CN106295522B (en) * | 2016-07-29 | 2019-09-10 | 武汉理工大学 | A kind of two-stage anti-fraud detection method based on multi-orientation Face and environmental information |
| WO2018040045A1 (en) * | 2016-08-31 | 2018-03-08 | 深圳前海达闼云端智能科技有限公司 | Monitoring method, apparatus and electronic device |
| CN106657544A (en) * | 2016-10-24 | 2017-05-10 | 广东欧珀移动通信有限公司 | Incoming call recording method and terminal equipment |
| CN106503687B (en) * | 2016-11-09 | 2019-04-05 | 合肥工业大学 | Merge the monitor video system for identifying figures and its method of face multi-angle feature |
| CN106503687A (en) * | 2016-11-09 | 2017-03-15 | 合肥工业大学 | The monitor video system for identifying figures of fusion face multi-angle feature and its method |
| CN107633244A (en) * | 2017-11-13 | 2018-01-26 | 姜茂清 | Human body direct picture harvester, the image capturing system comprising it and its application |
| WO2019223313A1 (en) * | 2018-05-22 | 2019-11-28 | 深圳云天励飞技术有限公司 | Personnel file establishment method and apparatus |
| CN109034178A (en) * | 2018-05-28 | 2018-12-18 | 北京文香信息技术有限公司 | A kind of demographic method based on face characteristic array |
| CN109214324A (en) * | 2018-08-27 | 2019-01-15 | 曜科智能科技（上海）有限公司 | Most face image output method and output system based on polyphaser array |
| CN113327332A (en) * | 2018-09-29 | 2021-08-31 | 爱笔（北京）智能科技有限公司 | Biological information-based ticket processing method and system, server and client |
| CN110781710A (en) * | 2018-12-17 | 2020-02-11 | 北京嘀嘀无限科技发展有限公司 | Target object clustering method and device |
| CN110338759B (en) * | 2019-06-27 | 2020-06-09 | 嘉兴深拓科技有限公司 | Facial pain expression data acquisition method |
| CN110338759A (en) * | 2019-06-27 | 2019-10-18 | 嘉兴深拓科技有限公司 | A kind of front pain expression data acquisition method |
| CN111539911A (en) * | 2020-03-23 | 2020-08-14 | 中国科学院自动化研究所 | Mouth breathing face recognition method, device and storage medium |
| CN111444822A (en) * | 2020-03-24 | 2020-07-24 | 北京奇艺世纪科技有限公司 | Object recognition method and apparatus, storage medium, and electronic apparatus |
| CN111444822B (en) * | 2020-03-24 | 2024-02-06 | 北京奇艺世纪科技有限公司 | Object recognition method and device, storage medium and electronic device |
| CN112001360A (en) * | 2020-09-09 | 2020-11-27 | 深圳中神电子科技有限公司 | Face recognition monitoring system based on intelligent adjustment |
| CN112001360B (en) * | 2020-09-09 | 2021-06-04 | 深圳市集互共享科技有限公司 | Face recognition monitoring system based on intelligent adjustment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102609695A (en) | Method and system for recognizing human face from multiple angles | |
| WO2021088300A1 (en) | Rgb-d multi-mode fusion personnel detection method based on asymmetric double-stream network | |
| US10204262B2 (en) | Infrared imaging recognition enhanced by 3D verification | |
| CN103824070B (en) | A kind of rapid pedestrian detection method based on computer vision | |
| CN108027874A (en) | Use the security system based on computer vision of depth camera | |
| CN103049084B (en) | A kind of electronic equipment and method thereof that can adjust display direction according to face direction | |
| CN104021538B (en) | Object positioning method and device | |
| CN105979210A (en) | Pedestrian identification system based on multi-ball multi-gun camera array | |
| CN103020606B (en) | Pedestrian detection method based on spatio-temporal context information | |
| US10789495B2 (en) | System and method for 1D root association providing sparsity guarantee in image data | |
| CN102930261A (en) | Face snapshot recognition method | |
| CN205068764U (en) | Face recognition system based on pedestrian's control of making a dash across red light | |
| De Oliveira et al. | People reidentification in a camera network | |
| Iwashita et al. | Gait identification using invisible shadows: robustness to appearance changes | |
| US20120162412A1 (en) | Image matting apparatus using multiple cameras and method of generating alpha maps | |
| WO2021248564A1 (en) | Panoramic big data application monitoring and control system | |
| CN111465937A (en) | Face detection and recognition method using light field camera system | |
| CN107025435B (en) | Face recognition processing method and system | |
| Jung et al. | Real-time estimation of 3D scene geometry from a single image | |
| DE102014222972A1 (en) | Image processing method, presence detector and lighting system | |
| Wang et al. | Image orientation detection with integrated human perception cues (or which way is up) | |
| CN116311041A (en) | Regional interest evaluation method, medium and device for intelligent stadium | |
| Wong et al. | Wireless webcam based omnidirectional health care surveillance system | |
| CN202472689U (en) | Face posture detection system | |
| Jagadeesh et al. | An conceptual view of an iris-biometric identification system canny edge detection 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 | ||
| C53 | Correction of patent of invention or patent application | ||
| CB03 | Change of inventor or designer information |
Inventor after: Hua Jiaobao Inventor after: Wu Minwei Inventor after: Liu Qifeng Inventor after: Qin Han Inventor after: Zhang Yaojun Inventor before: Hua Jiaobao Inventor before: Wu Minwei Inventor before: Liu Qifeng Inventor before: Qin Han |
|
| COR | Change of bibliographic data |
Free format text: CORRECT: INVENTOR; FROM: HUA JIAOBAO WU MINWEI LIU QIFENG QIN HAN TO: HUA JIAOBAO WU MINWEI LIU QIFENG QIN HAN ZHANG YAOJUN |
|
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20120725 |