CN104778462A - Face recognition method and device - Google Patents
Face recognition method and device Download PDFInfo
- Publication number
- CN104778462A CN104778462A CN201510208600.1A CN201510208600A CN104778462A CN 104778462 A CN104778462 A CN 104778462A CN 201510208600 A CN201510208600 A CN 201510208600A CN 104778462 A CN104778462 A CN 104778462A
- Authority
- CN
- China
- Prior art keywords
- image
- facial image
- face
- identified
- projection
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Collating Specific Patterns (AREA)
Abstract
The invention provides a face recognition method and device based on a mobile terminal. The method includes the steps of obtaining an image to be recognized, detecting a face image in the image to be recognized through a preset method, projecting the face image in characteristic face space, and recognizing the face image according to the comparison result of the projection coefficient obtained through projection and a projection coefficient library. By projecting the face image in the characteristic face space and recognizing the face image according to the comparison result of the projection coefficient obtained through projection and the projection coefficient library, the face recognition method and device can be widely applied to various intelligent terminals, and can be widely applied to the public security field and the mobile payment field.
Description
Technical field
The present invention relates to a kind of face identification method and device, belong to mobile technical field of biometric identification.
Background technology
Identity authentication is a part indispensable in the middle of people live, and along with the development of science and technology, the method for identity authentication is also following science and technology together to improve.The people that existing identity identifying method can be remembered at one's side by the feature such as appearance, sound, but these methods can produce memory error as time goes by.Therefore, utilize computing machine to carry out bio-identification and become one of hot research problem.Living things feature recognition comprises: the methods such as fingerprint recognition, handwriting signature identification, voice recognition, retina identification, iris recognition, identify genes and recognition of face.Face identification method analyzes facial image by computing machine in fact exactly, according to features different between people's face, carrys out a technology of identification identity.
Face identification method a kind ofly obtains facial image by camera and to be analyzed it by computing machine and pick out a kind of technology of facial image identity.According to the situation of research now, existing face identification method can comprise step as shown in Figure 1 usually.
Face recognition technology, compared with other biological identification technology, it is advantageous that: collection apparatus is convenient, does not need target to be placed on by finger on Fingerprint Identification Unit or by retina and aims at iris scan device, only need by camera collection picture; Non-invasion, this might not refer to its acquisition image of taking pictures in the ignorant situation of the people of collected image, finds oneself to be photographed, but as long as anything will be caused to affect on it, just can't dislike this even if also comprise target person.Therefore, face recognition technology has development prospect very widely, has very important researching value in field of biological recognition.
But face recognition technology also exists a lot of difficulty at present to be had to be solved.First based on recognition of face itself, it is unlike the recognition technology having notable difference between numeral or those individualities of Text region, and the face shape between face with face is more similar on the whole with distribution.This similarity just, for human face detection tech provides great convenience---can easier position of orienting face place in picture, but this but considerably increases the difficulty of face recognition technology.And face is except similarity, also have the expression shape change of face, for this problem, just need to try not when Feature Selection to choose as mouth, this face organ that can change of nose.Except the problem of face aspect, also has intensity of illumination, the problem of the target image change that the difference of lighting angle also can cause.
At present, this area is less for the technical research realizing recognition of face on mobile terminals, although the face recognition technology of existing such as " face payment " function, its range of application is more single, and the accuracy in plurality of human faces identification is poor.
Summary of the invention
The present invention is the problem poor by the accuracy of mobile terminal to recognition of face solving the existence of existing face recognition technology, and then provides a kind of face identification method based on mobile terminal and device.For this reason, the invention provides following technical scheme:
Based on a face identification method for mobile terminal, comprising:
Obtain image to be identified;
The facial image in described image to be identified is detected by preordering method;
Described facial image is projected in eigenface space, and according to the described projection coefficient of projection acquisition and the comparative result in projection coefficient storehouse, described facial image is identified.
Based on a face identification device for mobile terminal, comprising:
Image collection module, for obtaining image to be identified;
Image detection module, for detecting the facial image in described image to be identified by preordering method;
Picture recognition module, for being projected in eigenface space by described facial image, and identifies described facial image according to the described projection coefficient of projection acquisition and the comparative result in projection coefficient storehouse.
Technical scheme provided by the invention is by projecting facial image in eigenface space, and according to the projection coefficient of projection acquisition and the comparative result in projection coefficient storehouse, facial image is identified, various intelligent terminal can be widely used in, can be widely used in public safety field, mobile payment field.
Accompanying drawing explanation
In order to be illustrated more clearly in the technical scheme of the embodiment of the present invention, below the accompanying drawing used required in describing embodiment is briefly described, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is the schematic flow sheet of existing face identification method;
Fig. 2 is the schematic flow sheet of the face identification method based on mobile terminal that the specific embodiment of the present invention provides;
Fig. 3 is the schematic flow sheet of the face identification method of the Based PC A method that the specific embodiment of the present invention provides;
Fig. 4 is the schematic flow sheet of the face identification method after the optimization that provides of the specific embodiment of the present invention;
Fig. 5 is the schematic flow sheet of the face identification method based on Android mobile terminal that the specific embodiment of the present invention provides;
Fig. 6 is the structural representation of the face identification device based on mobile terminal that the specific embodiment of the present invention provides.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.
This embodiment provides a kind of face identification method based on mobile terminal, as shown in Figure 2, comprising:
Step 21, obtains image to be identified.
Wherein, this image to be identified obtains by the first-class image acquisition equipment of the shooting in mobile terminal, and requires that the sharpness of facial image at least wants can be relatively easy to tell the position at face place.
Existing mobile terminal mainly comprises iOS and Android two kinds of operating systems.IOS system is closed, although such system is stablized, its source code is nonopen, and developer wants the application program researching and developing iOS system, is cumbersome.And compared to iOS system, the opening of Android platform is more rising.Application program Java in Android operation system writes, if developer wants to use C/C++ language compilation application program in android system, can install Android NDK to realize.Android system itself also supports the form of music as multiple in MP3, WAV, MP4,3PG, JPEG, PNG etc., video, picture, convenient time this makes developer's developing application.
Step 22, detects the facial image in described image to be identified by preordering method.
Wherein, the detecting step of facial image mainly judges whether have face to exist in the image got, if can find face in the picture, then can utilize the geometrical normalization in the normalized of image that the position at face place is found out.
Step 23, projects described facial image in eigenface space, and identifies described facial image according to the described projection coefficient of projection acquisition and the comparative result in projection coefficient storehouse.
Wherein, recognition of face is higher to image request, therefore can carry out pre-service to image.Traditional preprocess method is the geometric manipulations way by rotation, convergent-divergent, adds this optical processing way of gray scale normalization.The present embodiment can adopt the method extracting facial image feature (PCA).
This PCA method utilizes Karhunen-Loeve transformation to extract the principal ingredient of face, by retaining the larger orthogonal basis of dimensional images space eigenvalues, comes constitutive characteristic face space.The problem of a higher dimensional space is put into a lower dimensional space to solve.When by the time identifying, obtain a new facial image as test pattern, test pattern is projected in this space, obtain one group of projection coefficient, again other face in this group coefficient and face database is compared, if similarity is higher, then illustrate that target face belongs to this face database, otherwise then illustrate that target face does not belong to this storehouse.
PCA technology mainly carries out the technology of dimensionality reduction in the picture in recognition of face to the feature space of face.By asking one group of orthogonal vector to original face, utilize this group vector to form new face characteristic space, and then constitute eigenface.In eigenface, still retain the face information in this original image.Use the algorithm concrete steps of PCA technology as shown in Figure 3, comprise loading training image, described eigenface space is generated by the facial image training of PCA method to the predetermined quantity taken in, and set up described projection coefficient storehouse by described eigenface space, by this eigenface space, training set is mapped to PCA subspace, and the information of preserving is face data file, by image mapped to be identified to PCA subspace, treat recognition image by the Euclidean distance of the pre-determined factor in projection coefficient and library of factors to identify, calculate and obtain the storehouse image maximum with images match rate to be identified, thus acquisition recognition result.
In an alternate embodiment of the present invention, generate described eigenface space by the facial image training of PCA method to the predetermined quantity taken in comprise: described eigenface spatial division is become multiple subspace by the face feature according to face, and the subspace belonging to described facial image is detected respectively by PCA method in each described subspace.
For the division of multiple subspace, the present embodiment can when setting up face space, execution step as shown in Figure 4, is divided into beard, without beard, four sub spaces of wearing glasses and not wearing glasses, every sub spaces carries out PCA feature extraction respectively.When treating the facial image in recognition image and identifying, first judge to belong to which class subspace, carry out projection coefficient corresponding to proper vector with wherein qualified subspace to contrast, thus decrease the operand of system, decrease recognition time, add discrimination, achieve the optimization of algorithm and link.
The programming language that android system storehouse is supported is Java language, therefore needs first to build Java Runtime Environment.After putting up Java Runtime Environment, next Android developing instrument is installed.More common Android developing instrument is Eclipse software at present.Eclipse installs ADT (Android Development Tools) kit.
Based on the mobile terminal of Android platform, apply PCA method and absorption facial image is trained, morphogenesis characters face space, and set up projection coefficient storehouse.Take in the face needing to identify, project in eigenface space, in projection coefficient and projection coefficient storehouse, data compare, and by judging Euclidean distance, identify this face.
Android platform uses PCA method, to the training of input facial image, carries out dimension-reduction treatment, be beneficial to and accelerate training and recognition speed.For the problem of Android platform arithmetic capability deficiency, treat recognition image by preordering method and compress, and by be applicable to image procossing OpenCV storehouse, thus effectively shorten the development time.
Below by specific embodiment, the face identification method based on mobile terminal that the present invention proposes is described in detail.
As shown in Figure 5, the main interface of the application program based on the face identification method of mobile terminal includes three buttons, is " interpolation face ", " emptying face " and " face verification " respectively.
After entering into main interface, according to the flow process of recognition of face, first can use training process, after clicking " interpolation face ", can enter and add face interface, before typing facial image, first input the ID of this sample.
After having inputted sample ID, click and determine button, start typing facial image.
In order to improve the travelling speed of discrimination and application program, when the sample that typing is new, can Select input five facial images as sample space, after adopting consecutive click chemical reaction screen five times, can the main interface of auto-returned, so far training step operation terminates.
Next be the operating process of identifying.Click " face verification " button, enter into recognition of face interface.
In identifying, if system detects face in the picture, with the line of blueness, face place place can be marked, now should ensure the stable of mobile terminal, so that system can identify face smoothly as far as possible.After successfully identifying, the ID of target face can be returned to user by application program
In order to facial image in anti-locking system too much causes arithmetic speed to decline, " emptying face " button can be clicked, facial image application program preserved in an operating system and fundamental function full scale clearance.
In test process, mainly for the test of face recognition software correct recognition rata, test group number is five people.First be training, five people everyone take five photos as training sample.The expression of the facial image of everyone typing, expression have certain change, and the final facial image obtained can be passed through normalized.After training link terminates, described facial image is projected in eigenface space, and according to the projection coefficient of projection acquisition and the comparative result in projection coefficient storehouse, described facial image is identified.Through test, when test number reaches 5 people, software can ensure that discrimination is 100%.
This embodiment additionally provides a kind of face identification device based on mobile terminal, as shown in Figure 6, comprising:
Image collection module 61, for obtaining image to be identified;
Image detection module 62, for detecting the facial image in described image to be identified by preordering method;
Picture recognition module 63, for being projected in eigenface space by described facial image, and identifies described facial image according to the described projection coefficient of projection acquisition and the comparative result in projection coefficient storehouse.
Optionally, comprise in picture recognition module 63:
Eigenface space training submodule, for generating described eigenface space by the facial image training of PCA method to the predetermined quantity taken in, and sets up described projection coefficient storehouse by described eigenface space.
Optionally, also comprise in picture recognition module 63:
Subspace partition submodule, for the face feature according to face, described eigenface spatial division is become multiple subspace, the subspace belonging to described facial image is detected respectively by PCA method in each described subspace.
Optionally, can also comprise in picture recognition module 63:
Recognin module, for identifying described facial image by the Euclidean distance of the pre-determined factor in described projection coefficient and described library of factors.
Optionally, also should comprise based on the face identification device of mobile terminal:
Compression module, for compressing described image to be identified by preordering method.
The specific implementation based on each module in the face identification device of mobile terminal described in the present embodiment is described in detail in aforesaid embodiment of the method, thus this no longer apply state.
Adopt the technical scheme that this embodiment provides, by facial image is projected in eigenface space, and according to the projection coefficient of projection acquisition and the comparative result in projection coefficient storehouse, facial image is identified, various intelligent terminal can be widely used in, can be widely used in public safety field, mobile payment field.
The above; be only the present invention's preferably embodiment; but protection scope of the present invention is not limited thereto; anyly be familiar with those skilled in the art in the technical scope that the embodiment of the present invention discloses; the change that can expect easily or replacement, all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claim.
Claims (10)
1. based on a face identification method for mobile terminal, it is characterized in that, comprising:
Obtain image to be identified;
The facial image in described image to be identified is detected by preordering method;
Described facial image is projected in eigenface space, and according to the described projection coefficient of projection acquisition and the comparative result in projection coefficient storehouse, described facial image is identified.
2. method according to claim 1, is characterized in that, described facial image is carried out projection in eigenface space and comprises:
Generate described eigenface space by the facial image training of PCA method to the predetermined quantity taken in, and set up described projection coefficient storehouse by described eigenface space.
3. method according to claim 2, is characterized in that, generates described eigenface space comprise by the facial image training of PCA method to the predetermined quantity taken in:
Described eigenface spatial division is become multiple subspace by the face feature according to face, and the subspace belonging to described facial image is detected respectively by PCA method in each described subspace.
4. method according to claim 1, is characterized in that, the projection coefficient obtained according to described projection and the comparative result in projection coefficient storehouse carry out identification to described facial image and comprise:
By the Euclidean distance of the pre-determined factor in described projection coefficient and described library of factors, described facial image is identified.
5. the method according to Claims 1-4 any one, is characterized in that, after acquisition image to be identified, described method also comprises:
By preordering method, described image to be identified is compressed.
6. based on a face identification device for mobile terminal, it is characterized in that, comprising:
Image collection module, for obtaining image to be identified;
Image detection module, for detecting the facial image in described image to be identified by preordering method;
Picture recognition module, for being projected in eigenface space by described facial image, and identifies described facial image according to the described projection coefficient of projection acquisition and the comparative result in projection coefficient storehouse.
7. device according to claim 6, is characterized in that, comprises in described picture recognition module:
Eigenface space training submodule, for generating described eigenface space by the facial image training of PCA method to the predetermined quantity taken in, and sets up described projection coefficient storehouse by described eigenface space.
8. device according to claim 7, is characterized in that, also comprises in described picture recognition module:
Subspace partition submodule, for the face feature according to face, described eigenface spatial division is become multiple subspace, the subspace belonging to described facial image is detected respectively by PCA method in each described subspace.
9. device according to claim 6, is characterized in that, also comprises in described picture recognition module:
Recognin module, for identifying described facial image by the Euclidean distance of the pre-determined factor in described projection coefficient and described library of factors.
10. the device according to claim 6 to 9 any one, is characterized in that, after acquisition image to be identified, described method also comprises:
By preordering method, described image to be identified is compressed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510208600.1A CN104778462A (en) | 2015-04-28 | 2015-04-28 | Face recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510208600.1A CN104778462A (en) | 2015-04-28 | 2015-04-28 | Face recognition method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104778462A true CN104778462A (en) | 2015-07-15 |
Family
ID=53619915
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510208600.1A Pending CN104778462A (en) | 2015-04-28 | 2015-04-28 | Face recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104778462A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018040028A1 (en) * | 2016-08-31 | 2018-03-08 | 张北江 | Method and system for recognizing face in video |
| CN113177466A (en) * | 2021-04-27 | 2021-07-27 | 北京百度网讯科技有限公司 | Identity recognition method and device based on face image, electronic equipment and medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020150280A1 (en) * | 2000-12-04 | 2002-10-17 | Pingshan Li | Face detection under varying rotation |
| US7254257B2 (en) * | 2002-03-04 | 2007-08-07 | Samsung Electronics Co., Ltd. | Method and apparatus of recognizing face using component-based 2nd-order principal component analysis (PCA)/independent component analysis (ICA) |
| CN101226591A (en) * | 2008-01-31 | 2008-07-23 | 上海交通大学 | Personal identification method based on mobile phone pick-up head combining with human face recognition technique |
| CN104143079A (en) * | 2013-05-10 | 2014-11-12 | 腾讯科技（深圳）有限公司 | Method and system for face attribute recognition |
-
2015
- 2015-04-28 CN CN201510208600.1A patent/CN104778462A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020150280A1 (en) * | 2000-12-04 | 2002-10-17 | Pingshan Li | Face detection under varying rotation |
| US7254257B2 (en) * | 2002-03-04 | 2007-08-07 | Samsung Electronics Co., Ltd. | Method and apparatus of recognizing face using component-based 2nd-order principal component analysis (PCA)/independent component analysis (ICA) |
| CN101226591A (en) * | 2008-01-31 | 2008-07-23 | 上海交通大学 | Personal identification method based on mobile phone pick-up head combining with human face recognition technique |
| CN104143079A (en) * | 2013-05-10 | 2014-11-12 | 腾讯科技（深圳）有限公司 | Method and system for face attribute recognition |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018040028A1 (en) * | 2016-08-31 | 2018-03-08 | 张北江 | Method and system for recognizing face in video |
| CN113177466A (en) * | 2021-04-27 | 2021-07-27 | 北京百度网讯科技有限公司 | Identity recognition method and device based on face image, electronic equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10339402B2 (en) | Method and apparatus for liveness detection | |
| Chan et al. | Face liveness detection using a flash against 2D spoofing attack | |
| WO2016127437A1 (en) | Live body face verification method and system, and computer program product | |
| KR102288302B1 (en) | Authentication method and authentication apparatus using infrared ray(ir) image | |
| US20120320181A1 (en) | Apparatus and method for security using authentication of face | |
| US20150154394A1 (en) | Gesture controlled login | |
| JP6798798B2 (en) | Method and device for updating data for user authentication | |
| KR20170046448A (en) | Method and device for complex authentication | |
| US11663853B2 (en) | Iris authentication device, iris authentication method, and recording medium | |
| US11704937B2 (en) | Iris authentication device, iris authentication method and recording medium | |
| CN112825128A (en) | Method and apparatus for liveness testing and/or biometric verification | |
| KR20170045813A (en) | Detecting method and apparatus of biometrics region for user authentication | |
| WO2016149985A1 (en) | Direction control method, direction control system and terminal | |
| CN111382666A (en) | Device and method with user authentication | |
| KR102205495B1 (en) | Method and apparatus for recognizing finger print | |
| KR20120026232A (en) | Method and apparatus for generating an objected descriptor using extended curvature gabor filter | |
| RU2608001C2 (en) | System and method for biometric behavior context-based human recognition | |
| CN104778462A (en) | Face recognition method and device | |
| KR20170003361A (en) | Method and apparatus for recognizing finger print | |
| US10133471B2 (en) | Biometric shortcuts | |
| KR20070118806A (en) | Method of detecting face for embedded system | |
| KR101286750B1 (en) | Password estimation system using gesture. | |
| Ouanan et al. | Implementation and optimization of face detection frameworkbased on OpenCV library on mobile platforms using Davinci’s technology | |
| Jonsson et al. | Learning salient features for real-time face verification | |
| Dhope et al. | Face Recognition Based Attendance System using SIFT Algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150715 |