CN105512655A - Face recognition method and face recognition device - Google Patents
Face recognition method and face recognition device Download PDFInfo
- Publication number
- CN105512655A CN105512655A CN201610121059.5A CN201610121059A CN105512655A CN 105512655 A CN105512655 A CN 105512655A CN 201610121059 A CN201610121059 A CN 201610121059A CN 105512655 A CN105512655 A CN 105512655A
- Authority
- CN
- China
- Prior art keywords
- posture
- image
- face
- model
- facial image
- 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.)
- Withdrawn
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
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
Abstract
The invention provides a face recognition method and a face recognition device. The face recognition method includes the steps of acquiring a face image; acquiring face illumination intensity; preprocessing the face image according to the illumination intensity; extracting face features from the preprocessed face image; comparing the extracted face features with all face templates to determine whether face recognition is passed or not, and reading identity information specified by a recognized person when face recognition is not passed after the extracted face features are compared with all the face templates; acquiring a specified face template according to the identity information specified by the recognized person; comparing the extracted face features with the specified face template to determine whether face recognition is passed or not. By the face recognition method, misrecognition and recognition rejection can be reduced, recognition success rate can be increased, human intervention is avoided and recognition efficiency is improved.
Description
Technical field
Of the present inventionly relate to image procossing, and more specifically relate to method, equipment and the computer program determining one or more object and form, be specifically related to face identification method and face identification device
Background technology
Modem communications era brings the wired of huge expansion and wireless network.Computer network, TV network and telephone network are just experiencing unprecedented technological transformation enlarging, and consumer demand have stimulated.Wireless and mobile network's technology meets relevant consumption demand, provides more dirigibility and direct information transmission simultaneously.
Current and following network technology continues facilitate the information transmission of ease for use and facilitate user.Because the present ubiquitous institute has age of electronic communication equipment all utilizes electronic equipment to communicate with other people or contact person with the people of level of education, receive service and/or shared information, media and other guide.One side is that the information transmission of increase in demand ease for use relates to execution image processing services.In this respect, such as, the one or more objects improved in the image determining a posture exactly of reliability or video can improve image procossing.
At present, the posture posture that may perform as detected object is based on object-detection device.Although traditional posture can detect and object of classification given pose based on object-detection device, the chance of traditional posture detects a fault according to object-detection device may higher than desirable.This is normally because the training of a posture sorter is usually by using identical posture, and the change in sample may be that huge many samples and/or feature selecting may not be enough powerful due to plantation in violation of rules and regulations, with a posture of representative object effectively.
Especially, traditional posture may run into the face with border based on object-detection device, and in the pose detection of the example of erroneous association, Problems existing is due to the detection of process, the unlawful practice of plantation aspect, etc.Such as, by utilizing traditional posture based on detector, form that discriminate against usually may less in border with two.Such as, the face at the edge of image may more difficult detection.Therefore, form the face in these borders, usually utilize the pose metrical error of traditional method testing result.In addition, tradition pendulum based on detector may run into mistake due to testing process.Such as, if a scanning window surrounds the front face of part, scanning window in instances should enclose full face, based on detector, tradition posture is usually determined that face detects and is corresponded to a half side-view posture (such as, driftage <45 degree) and not in the face of a basic position, because the view in window is close to the configuration file of half side-view than front.This may be the intrinsic property due to window scanning.In addition, utilize conventional posture to detect based on pose in detector, mistake may cause the irregular image of one or more cutting.
Summary of the invention
For the deficiencies in the prior art, the invention provides a kind of face identification method and face identification device.
To achieve these goals, the technical solution adopted for the present invention to solve the technical problems is: face identification method, comprising: receive at least one face; The multiple alternative facial image of the Computer image genration detected, based on the image detected, wherein generates the image detected of the pixel that candidate's face image changes based on one or more row or column; Analyze alternative facial image determines one or more posture relative sections at least one position or direction alternative facial image separately based at least one models of data; And determine that the image of at least one face corresponds to the one or more projects of a posture part based on the determined alternative face image data of model corresponding during the posture transmitted.
The method that the present invention further provides, an one model at least comprises multiple submodel, the posture that each model is corresponding respective, the posture of a diversification; The alternative facial image of described analysis comprises further analyzes each alternative facial image, to determine whether that alternative facial image passes through as each model; To settle the standard and the described image determining at least one face comprises for each model further corresponding to a posture, determine the facial image of candidate, by the total degree of standard.
The method that the present invention further provides, what at least one the face image wherein determined corresponded to that posture comprises the image detected determining at least one face corresponds to corresponding model, the posture of wherein maximum candidate's facial image Transfer Standards posture separately.
The method that the present invention further provides, wherein at least one model comprises multiple model, each correspondence posture separately, wherein the graphical analysis of at least one face comprises based on multiple models of at least one model of data the method that analysis part forms further based on the image detected of each model of data: be each model, determine whether that the image detected is by respective posture; One or more confidence score that at least one condition and calculating are associated by the conditional response of respective posture at the image determining to detect with each model.
The method that the present invention further provides, wherein at least one model comprises multiple model, each corresponding respective posture, multiple models wherein based at least one model analyzing alternative face image data comprise each alternative facial image of analysis, to determine whether the method that each model that each candidate face image transfer standard is set up forms further: for each model, determine the facial image of each candidate, by the confidence score of standard.
The method that the present invention further provides, it forms further: add the corresponding model of each confidence score to obtain multiple total confidence score, the pattern of the model that each total confidence score is corresponding respective; At least one the face image wherein determined corresponds to a further posture, comprises that to determine that the image of an at least one face of posture corresponds to the highest total confidence score of total confidence score that corresponding model is defined as forming be respective posture.
The method that the present invention further provides, wherein at least one model comprises the canonical correlation analysis model of corresponding multiple posture, each posture distributes corresponding label, wherein analyzes alternative facial image and comprises based at least one models of data and analyze the image detected and determine whether that the image detected corresponds to a posture and the method that the data of the data associated form further with attitude: the image that corresponding posture label distribution is determined to detect to middle and corresponding posture; Associated response detect image wherein the image of at least one face correspond to posture and determine that a posture is in the face of relating to corresponding posture at least partly based on the value of distributing labels based on determining section.
The method that the present invention further provides, wherein at least one model comprises the canonical correlation analysis model of corresponding multiple posture, each posture distributes a corresponding multiple label, the method of label forms further: by respective label, and label distribution gives the resolution of each candidate and the facial image of at least one posture of relative section.
Face identification device, comprising: at least one processor; Comprise computer program code with at least one internal memory, use at least one processor, configure the equipment caused, at least perform following steps: receive at least one face; The multiple alternative facial image of the Computer image genration detected is based on image generation being detected, and one of them candidate face image is the image detected of the pixel based on one or more row or column; Transfer analysis is based at least one position of a kind of pattern-recognition one or more posture relevant portion of data or direction alternative facial image separately; Determine that the image of at least one face corresponds to posture part one or more projects based on the determined alternative face image data of model corresponding during the posture transmitted at alternative facial image.
Beneficial effect of the present invention: face identification method provided by the invention and face identification device, after extraction face characteristic, this face characteristic and whole face template are compared, if passed through, then recognition of face terminates, if do not passed through, obtain the appointment face template of identified person according to identified person's identity information, and extracted face characteristic and the face template of specifying are compared, thus determine whether recognition of face is passed through.Owing to having carried out supplementary recognition of face, thus can reduce and identify by mistake and refuse to identify, improve recognition success rate, avoid human intervention simultaneously, improve recognition efficiency.
Embodiment
Face identification method, comprising: receive at least one face; The multiple alternative facial image of the Computer image genration detected, based on the image detected, wherein generates the image detected of the pixel that candidate's face image changes based on one or more row or column; Analyze alternative facial image determines one or more posture relative sections at least one position or direction alternative facial image separately based at least one models of data; And determine that the image of at least one face corresponds to the one or more projects of a posture part based on the determined alternative face image data of model corresponding during the posture transmitted.
Described face identification method, an one model at least comprises multiple submodel, the posture that each model is corresponding respective, the posture of a diversification; The alternative facial image of described analysis comprises further analyzes each alternative facial image, to determine whether that alternative facial image passes through as each model; To settle the standard and the described image determining at least one face comprises for each model further corresponding to a posture, determine the facial image of candidate, by the total degree of standard.
Described face identification method, what at least one the face image wherein determined corresponded to that posture comprises the image detected determining at least one face corresponds to corresponding model, the posture of wherein maximum candidate's facial image Transfer Standards posture separately.
Described face identification method, wherein at least one model comprises multiple model, each correspondence posture separately, wherein the graphical analysis of at least one face comprises based on multiple models of at least one model of data the method that analysis part forms further based on the image detected of each model of data: be each model, determine whether that the image detected is by respective posture; One or more confidence score that at least one condition and calculating are associated by the conditional response of respective posture at the image determining to detect with each model.
Described face identification method, wherein at least one model comprises multiple model, each corresponding respective posture, multiple models wherein based at least one model analyzing alternative face image data comprise each alternative facial image of analysis, to determine whether the method that each model that each candidate face image transfer standard is set up forms further: for each model, determine the facial image of each candidate, by the confidence score of standard.
Described face identification method, it forms further: add the corresponding model of each confidence score to obtain multiple total confidence score, the pattern of the model that each total confidence score is corresponding respective; At least one the face image wherein determined corresponds to a further posture, comprises that to determine that the image of an at least one face of posture corresponds to the highest total confidence score of total confidence score that corresponding model is defined as forming be respective posture.
Described face identification method, wherein at least one model comprises the canonical correlation analysis model of corresponding multiple posture, each posture distributes corresponding label, wherein analyzes alternative facial image and comprises based at least one models of data and analyze the image detected and determine whether that the image detected corresponds to a posture and the method that the data of the data associated form further with attitude: the image that corresponding posture label distribution is determined to detect to middle and corresponding posture; Associated response detect image wherein the image of at least one face correspond to posture and determine that a posture is in the face of relating to corresponding posture at least partly based on the value of distributing labels based on determining section.
Described face identification method, wherein at least one model comprises the canonical correlation analysis model of corresponding multiple posture, each posture distributes a corresponding multiple label, the method of label forms further: by respective label, and label distribution gives the resolution of each candidate and the facial image of at least one posture of relative section.
Face identification device, comprising: at least one processor; Comprise computer program code with at least one internal memory, use at least one processor, configure the equipment caused, at least perform following steps: receive at least one face; The multiple alternative facial image of the Computer image genration detected is based on image generation being detected, and one of them candidate face image is the image detected of the pixel based on one or more row or column; Transfer analysis is based at least one position of a kind of pattern-recognition one or more posture relevant portion of data or direction alternative facial image separately; Determine that the image of at least one face corresponds to posture part one or more projects based on the determined alternative face image data of model corresponding during the posture transmitted at alternative facial image.
To those skilled in the art, obviously the invention is not restricted to the details of above-mentioned one exemplary embodiment, and when not deviating from spirit of the present invention or essential characteristic, the present invention can be realized in other specific forms.Therefore, no matter from which point, all should embodiment be regarded as exemplary, and be nonrestrictive, scope of the present invention is limited by claims instead of above-mentioned explanation, and all changes be therefore intended in the implication of the equivalency by dropping on claim and scope are included in the present invention.
In addition, be to be understood that, although this instructions is described according to embodiment, but not each embodiment only comprises an independently technical scheme, this narrating mode of instructions is only for clarity sake, those skilled in the art should by instructions integrally, and the technical scheme in each embodiment also through appropriately combined, can form other embodiments that it will be appreciated by those skilled in the art that.
Claims (9)
1. face identification method, is characterized in that, comprising: receive at least one face; The multiple alternative facial image of the Computer image genration detected, based on the image detected, wherein generates the image detected of the pixel that candidate's face image changes based on one or more row or column; Analyze alternative facial image determines one or more posture relative sections at least one position or direction alternative facial image separately based at least one models of data; And determine that the image of at least one face corresponds to the one or more projects of a posture part based on the determined alternative face image data of model corresponding during the posture transmitted.
2. method according to claim 1, is characterized in that, a model at least comprises multiple submodel, the posture that each model is corresponding respective, the posture of a diversification; The alternative facial image of described analysis comprises further analyzes each alternative facial image, to determine whether that alternative facial image passes through as each model; To settle the standard and the described image determining at least one face comprises for each model further corresponding to a posture, determine the facial image of candidate, by the total degree of standard.
3. method according to claim 1, it is characterized in that, what at least one the face image wherein determined corresponded to that posture comprises the image detected determining at least one face corresponds to corresponding model, the posture of wherein maximum candidate's facial image Transfer Standards posture separately.
4. method according to claim 1, it is characterized in that, wherein at least one model comprises multiple model, each correspondence posture separately, wherein the graphical analysis of at least one face comprises based on multiple models of at least one model of data the method that analysis part forms further based on the image detected of each model of data: be each model, determine whether that the image detected is by respective posture; One or more confidence score that at least one condition and calculating are associated by the conditional response of respective posture at the image determining to detect with each model.
5. method according to claim 1, it is characterized in that, wherein at least one model comprises multiple model, each corresponding respective posture, multiple models wherein based at least one model analyzing alternative face image data comprise each alternative facial image of analysis, to determine whether the method that each model that each candidate face image transfer standard is set up forms further: for each model, determine the facial image of each candidate, by the confidence score of standard.
6. method according to claim 1, is characterized in that, forms further: add the corresponding model of each confidence score to obtain multiple total confidence score, the pattern of the model that each total confidence score is corresponding respective; At least one the face image wherein determined corresponds to a further posture, comprises that to determine that the image of an at least one face of posture corresponds to the highest total confidence score of total confidence score that corresponding model is defined as forming be respective posture.
7. method according to claim 1, it is characterized in that, wherein at least one model comprises the canonical correlation analysis model of corresponding multiple posture, each posture distributes corresponding label, wherein analyzes alternative facial image and comprises based at least one models of data and analyze the image detected and determine whether that the image detected corresponds to a posture and the method that the data of the data associated form further with attitude: the image that corresponding posture label distribution is determined to detect to middle and corresponding posture; Associated response detect image wherein the image of at least one face correspond to posture and determine that a posture is in the face of relating to corresponding posture at least partly based on the value of distributing labels based on determining section.
8. method according to claim 1, it is characterized in that, wherein at least one model comprises the canonical correlation analysis model of corresponding multiple posture, each posture distributes a corresponding multiple label, the method of label forms further: by respective label, and label distribution gives the resolution of each candidate and the facial image of at least one posture of relative section.
9. face identification device, comprising: at least one processor; Comprise computer program code with at least one internal memory, use at least one processor, configure the equipment caused, at least perform following steps: receive at least one face; The multiple alternative facial image of the Computer image genration detected is based on image generation being detected, and one of them candidate face image is the image detected of the pixel based on one or more row or column; Transfer analysis is based at least one position of a kind of pattern-recognition one or more posture relevant portion of data or direction alternative facial image separately; Determine that the image of at least one face corresponds to posture part one or more projects based on the determined alternative face image data of model corresponding during the posture transmitted at alternative facial image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610121059.5A CN105512655A (en) | 2016-03-03 | 2016-03-03 | Face recognition method and face recognition device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610121059.5A CN105512655A (en) | 2016-03-03 | 2016-03-03 | Face recognition method and face recognition device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105512655A true CN105512655A (en) | 2016-04-20 |
Family
ID=55720622
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610121059.5A Withdrawn CN105512655A (en) | 2016-03-03 | 2016-03-03 | Face recognition method and face recognition device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105512655A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107657703A (en) * | 2017-10-25 | 2018-02-02 | 成都云凯软件有限责任公司 | A kind of community's discrepancy personnel management system based on Internet of Things cloud |
CN107784723A (en) * | 2017-10-25 | 2018-03-09 | 成都云凯软件有限责任公司 | Community's discrepancy personnel's management-control method based on Internet of Things cloud |
-
2016
- 2016-03-03 CN CN201610121059.5A patent/CN105512655A/en not_active Withdrawn
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107657703A (en) * | 2017-10-25 | 2018-02-02 | 成都云凯软件有限责任公司 | A kind of community's discrepancy personnel management system based on Internet of Things cloud |
CN107784723A (en) * | 2017-10-25 | 2018-03-09 | 成都云凯软件有限责任公司 | Community's discrepancy personnel's management-control method based on Internet of Things cloud |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10405052B2 (en) | Method and apparatus for identifying television channel information | |
CN111339990B (en) | Face recognition system and method based on dynamic update of face features | |
JP4543423B2 (en) | Method and apparatus for automatic object recognition and collation | |
CN111144366A (en) | Strange face clustering method based on joint face quality assessment | |
CN109376604B (en) | Age identification method and device based on human body posture | |
CN103618918A (en) | Method and device for controlling display of smart television | |
CN110119726A (en) | A kind of vehicle brand multi-angle recognition methods based on YOLOv3 model | |
CN110610127B (en) | Face recognition method and device, storage medium and electronic equipment | |
CN111091080A (en) | Face recognition method and system | |
Paul et al. | Extraction of facial feature points using cumulative histogram | |
Li et al. | Fast and effective text detection | |
CN110796039B (en) | Face flaw detection method and device, electronic equipment and storage medium | |
Lee et al. | Robust iris recognition baseline for the grand challenge | |
CN105550642B (en) | Gender identification method and system based on multiple dimensioned linear Differential Characteristics low-rank representation | |
CN105512655A (en) | Face recognition method and face recognition device | |
EP2659429B1 (en) | Methods, apparatuses and computer program products for efficiently recognizing faces of images associated with various illumination conditions | |
Fritz et al. | Attentive object detection using an information theoretic saliency measure | |
CN108537990A (en) | All-in-one machine cheats judgment method, device, equipment and computer readable storage medium | |
CN110222660B (en) | Signature authentication method and system based on dynamic and static feature fusion | |
Li et al. | Biometrics at a distance: issues, challenges, and prospects | |
CN111800428A (en) | Real-time statistical method and system for digital conference participation | |
CN108830217B (en) | Automatic signature distinguishing method based on fuzzy mean hash learning | |
Paul et al. | Automatic adaptive facial feature extraction using CDF analysis | |
CN105323143A (en) | Network information pushing method, apparatus and system based on instant message | |
Paul et al. | Extraction of facial feature points using cumulative distribution function by varying single threshold group |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WW01 | Invention patent application withdrawn after publication |
Application publication date: 20160420 |
|
WW01 | Invention patent application withdrawn after publication |