CN109711384A - A kind of face identification method based on depth convolutional neural networks - Google Patents
A kind of face identification method based on depth convolutional neural networks Download PDFInfo
- Publication number
- CN109711384A CN109711384A CN201910017928.3A CN201910017928A CN109711384A CN 109711384 A CN109711384 A CN 109711384A CN 201910017928 A CN201910017928 A CN 201910017928A CN 109711384 A CN109711384 A CN 109711384A
- Authority
- CN
- China
- Prior art keywords
- face
- indicate
- facial image
- ellipse
- convolutional neural
- 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
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 19
- 230000001815 facial effect Effects 0.000 claims abstract description 31
- 238000010586 diagram Methods 0.000 claims abstract description 20
- 238000003062 neural network model Methods 0.000 claims abstract description 15
- 239000011159 matrix material Substances 0.000 claims abstract description 11
- 239000000284 extract Substances 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims description 15
- 238000013528 artificial neural network Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 6
- 239000000203 mixture Substances 0.000 claims description 3
- 230000007613 environmental effect Effects 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 230000007717 exclusion Effects 0.000 description 1
- 238000003780 insertion Methods 0.000 description 1
- 230000037431 insertion Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
Abstract
Invention broadly provides a kind of face identification methods based on depth convolutional neural networks, including following content: obtaining original facial image using camera；Original facial image is input in MTCNN neural network model, the face face block diagram picture after face face is cut is obtained；The positive face face block diagram picture of face is done into inscribed ellipse, extracts the facial image in ellipse；Facial image in ellipse is input in LCNN neural network model, face characteristic matrix is obtained；Identification is compared with the eigenmatrix of the facial image in face template library in face characteristic matrix and obtains face result.The present invention provides a kind of face identification method based on depth convolutional neural networks, carries out background removal before the defeated depth convolutional neural networks of image, is extracted face part using elliptical form is done, and influence of the environmental factor to image content is discharged.
Description
Technical field:
The present invention relates to image procossing identification technology more particularly to a kind of recognition of face sides based on depth convolutional neural networks
Method.
Background technique:
In modern society, personal identification technology using omnipresent, wherein based on fingerprint, iris and face etc.
There are huge market demands in many areas for the identification technology of human body biological characteristics, such as: access control system, video monitoring, airport
Safety check and intelligent space etc..Although it is higher accurate that the authentication based on fingerprint and iris has than face recognition technology
Property and reliability, but recognition of face because with nature, close friend, to user's interference less, the advantages such as susceptible to user acceptance due to have it is wider
Wealthy application prospect.
Recognition of face is based on technologies such as Digital Image Processing, computer vision and machine learning, at computer
Reason technology carries out the process of analysis comparison to facial image in database.Currently, the method for recognition of face mainly utilizes depth
Convolutional neural networks complete to identify by convolution operation repeatedly, this method in actual use, often because of input
Image is mixed into background and excessively causes not preparing to identify, because the background that image can have a part when cutting be mixed into,
This mixed background will affect recognition of face as a result, causing accuracy of identification low.
Summary of the invention:
To solve the above-mentioned problems, the present invention provides a kind of face identification method based on depth convolutional neural networks, in image
Background removal is carried out before defeated depth convolutional neural networks, is extracted face part using elliptical form is done, environment is discharged
Influence of the factor to image content.
In order to achieve the above objectives, the technical scheme is that a kind of recognition of face based on depth convolutional neural networks
Method includes the following steps:
Step 1) obtains original facial image using camera；
Original facial image is input in MTCNN neural network model by step 2, obtains the people after face face is cut
Face face block diagram picture；
The positive face face block diagram picture of face is done inscribed ellipse by step 3), extracts the facial image in ellipse；
Facial image in ellipse is input in LCNN neural network model by step 4), obtains face characteristic matrix；
Face characteristic matrix is compared identification with the eigenmatrix of the facial image in face template library and obtains people by step 5)
Face result.
Preferably, handling to obtain the process of face face block diagram picture using MTCNN neural network model in the step 2
Include:
Step 2-1, candidate forms and boundary regressor are obtained using P-Net network, while candidate forms are carried out according to bounding box
Calibration recycles NMS method removal overlapping forms；
Step 2-2, the picture comprising candidate forms for determining P-Net network training in R-Net network, using bounding box to
Amount finely tunes candidate framework, recycles NMS method removal overlapping forms；
Step 2-3, candidate forms are being removed using O-Net network, while is showing five face key point positioning.
Preferably, the training of the MTCNN neural network model includes three parts, face and non-face classification,
Bounding box returns and face position point location, in which:
Face and non-face classified use cross entropy loss function determine:
；
Wherein,Indicate face probability,Indicate the true tag of background,Indicate " probability of prediction face " and " true
Whether face " degree of closeness,Be worth it is smaller indicate closer, the training objective of the part is to obtain minimum value min()；
Bounding box, which returns to calculate by Euclidean distance, returns loss:
；
Wherein,For the background coordination obtained by neural network forecast,For actual true background coordination,Indicate left
The four-tuple that upper angle, the upper right corner, length and width form,Indicate frame return Euclidean distance, be worth it is smaller, indicate predicted value with
True value is closer, and the training objective of the part is to obtain minimum value min()；
Face position point location minimizes the distance by calculating the Euclidean distance of neural network forecast coordinate and actual coordinate,
Formula are as follows:
；
Wherein,For the face position coordinate obtained by neural network forecast,For actual true face position
Coordinate is set,Indicate ten tuples of five points composition of face,Indicate that the Euclidean distance that frame returns, value are got over
It is small, indicate that predicted value and true value are closer, the training objective of the part is to obtain minimum value min()；
Above-mentioned three parts are integrated to obtain formula:
；
Wherein: N is the quantity of training program,The weight for indicating different task, in P-Net network and R-Net network,=1,=0.5,=0.5, in O-Net network,=1,=1,=0.5；It indicates
The true tag of sample type,,Indicate loss function.
Preferably, each pixel in the positive face face block diagram picture of face is recycled after ellipse is inscribed in the step 3), and
Judge that the pixel whether in ellipse, retains if in ellipse, if not ignoring, to obtain oval facial image.
Preferably, the human face data information in the step 5) in face template library shifts to an earlier date typing, and also interior when typing
Connect image zooming-out in oval progress ellipse.
Preferably, in the step 5), by the eigenmatrix of facial image in face characteristic matrix and face template library
Cosine similarity operation is done, the similarity value of image in original image and face template library is obtained.
Beneficial effect, a kind of face identification method based on depth convolutional neural networks that the present invention discloses have as follows
The utility model has the advantages that
By the inscribed ellipse of rectangular face face block diagram picture, the facial image for extracting face in ellipse carries out collection apparatus, can be effective
Excessive influence of the environmental factor to image content in rejection image avoids influence of the background of the picture of acquisition to recognition of face.
Detailed description of the invention:
Fig. 1 is MTCNN neural network cascade schematic diagram of the present invention；
Fig. 2 is the schematic diagram of the inscribed ellipses recognition elliptical image of the present invention.
Specific embodiment:
Technology of the invention is described further below with reference to attached drawing provided by the present invention:
A kind of disclosed face identification method based on depth convolutional neural networks, includes the following steps:
Step 1) obtains original facial image using camera；
Original facial image is input in MTCNN neural network model by step 2, obtains the people after face face is cut
Face face block diagram picture；
The positive face face block diagram picture of face is done inscribed ellipse by step 3), extracts the facial image in ellipse；
Facial image in ellipse is input in LCNN neural network model by step 4), obtains face characteristic matrix；
Face characteristic matrix is compared identification with the eigenmatrix of the facial image in face template library and obtains people by step 5)
Face result.
Above-mentioned steps are described in detail below, wherein the acquisition of the original facial image of step 1) can be smart phone
Or other smart machines are obtained；
As shown in Figure 1, handling to obtain the process packet of face face block diagram picture using MTCNN neural network model in the step 2
It includes:
Step 2-1, candidate forms and boundary regressor are obtained using P-Net network, while candidate forms are carried out according to bounding box
Calibration recycles NMS method removal overlapping forms；
Step 2-2, the picture comprising candidate forms for determining P-Net network training in R-Net network, using bounding box to
Amount finely tunes candidate framework, recycles NMS method removal overlapping forms；
Step 2-3, candidate forms are being removed using O-Net network, while is showing five face key point positioning.
The training process of MTCNN neural network model includes three parts, face and non-face classification, and bounding box returns
And face position point location, in which:
Face and non-face classified use cross entropy loss function determine:
；
Wherein,Indicate face probability,Indicate the true tag of background,Indicate " probability of prediction face " and " true
Whether face " degree of closeness,Be worth it is smaller indicate closer, the training objective of the part is to obtain minimum value min()；
Bounding box, which returns to calculate by Euclidean distance, returns loss:
；
Wherein,For the background coordination obtained by neural network forecast,For actual true background coordination,Indicate left
The four-tuple that upper angle, the upper right corner, length and width form,Indicate frame return Euclidean distance, be worth it is smaller, indicate predicted value with
True value is closer, and the training objective of the part is to obtain minimum value min()；
Face position point location minimizes the distance by calculating the Euclidean distance of neural network forecast coordinate and actual coordinate,
Formula are as follows:
；
Wherein,For the face position coordinate obtained by neural network forecast,For actual true face position
Coordinate is set,Indicate ten tuples of five points composition of face,Indicate that the Euclidean distance that frame returns, value are got over
It is small, indicate that predicted value and true value are closer, the training objective of the part is to obtain minimum value min()；
Above-mentioned three parts are integrated to obtain formula:
；
Wherein: N is the quantity of training program,The weight for indicating different task, in P-Net network and R-Net network,=1,=0.5,=0.5, in O-Net network,=1,=1,=0.5；It indicates
The true tag of sample type,,Indicate loss function.
What is got is original facial image by having been obtained after face are aligned face cutting among the processing of MTCNN
Rectangular face face block diagram picture, the background due to some when cutting out face is mixed into face face block diagram piece,
It will affect the score of face recognition result in practice, so we are just directly by face when generating face template
Picture do ellipse, only show the part of face in a template, effectively shadow of the factor of exclusion environment to the content of picture
It rings.It is the also oval processing of this process of use progress in typing certainly for the facial image in face template library, so template
Facial image in library is also only to show face partial content, can be known in this way to avoid background is reduced when comparing in the later period to face
Other influence.
In specific step 3), after the inscribed ellipse of face face block diagram picture, each picture in the positive face face block diagram picture of face is recycled
Vegetarian refreshments, and judge that the pixel whether in ellipse, retains if in ellipse, if not ignoring, to obtain oval people
Face image, wherein oval step is inscribed are as follows:
Using the top left corner apex of original picture as origin, X-axis is established horizontally to the right, establishes Y-axis straight down, passes through Face datection
Algorithm, if there is face in picture, then it is known that the top left co-ordinate of face rectangle frame is (x1, y1), lower right corner top
The coordinate of point is (x2, y2), and any coordinate less than 0 is all assigned a value of 0；
Elliptical long axis long=(y2-y1)/2 is calculated by this four coordinates；Short axle short=(x2-x1)/2,
The coordinate O (centerX, centerY) in the oval center of circle is calculated simultaneously, wherein: centerX=x1+short, centerY
= y1 + long；
Judge each pixel (m, n) whether in elliptical inside, formula with following formula are as follows:
If operating structure is less than 1.0, then indicating that the point in elliptical inside, retains pixel, otherwise just not in ellipse
Portion carries out pixel and ignores.
LCNN neural network model is used in the step 4), is the lightweight depth convolutional Neural with noise label
Network, for learning the compact insertion of the extensive human face data with much noise label, the model is by the change of maximum activation
Change each convolutional layer that (referred to as maximum Feature Mapping (MFM)) introduces CNN, it is arbitrarily convex with using many features to map linear approximation
The maximum output value of activation primitive is different, and MFM is realized by competitive relation.MFM not only can with burbling noise and information signal,
It can also play the role of feature selecting between two Feature Mappings, model proposes a kind of semantic Bootload, makes network
It predicts more consistent with noise label.
In the step 5), it is similar that face characteristic matrix to the eigenmatrix of facial image in face template library is done into cosine
Operation is spent, the similarity value of image in original image and face template library is obtained.(such as attendance checking system) in actual use,
A threshold value is set, if obtained similarity value is higher than threshold value, is judged as similar pictures, attendance success is assert, if low
In threshold value, then it is judged as different pictures, indicates attendance failure.
Technology contents and technical characteristic of the invention have revealed that as above, however those skilled in the art still may base
Make various replacements and modification without departing substantially from spirit of that invention in announcement of the invention, therefore, the scope of the present invention is answered unlimited
It in the revealed content of embodiment, and should include various without departing substantially from replacement and modification of the invention, and be present patent application right
It is required that being covered.
Claims (6)
1. a kind of face identification method based on depth convolutional neural networks, which comprises the steps of:
Step 1) obtains original facial image using camera；
Original facial image is input in MTCNN neural network model by step 2, obtains the people after face face is cut
Face face block diagram picture；
The positive face face block diagram picture of face is done inscribed ellipse by step 3), extracts the facial image in ellipse；
Facial image in ellipse is input in LCNN neural network model by step 4), obtains face characteristic matrix；
Face characteristic matrix is compared identification with the eigenmatrix of the facial image in face template library and obtains people by step 5)
Face result.
2. the face identification method according to claim 1 based on depth convolutional neural networks, it is characterised in that: the step
It is rapid 2) in handle to obtain the process of face face block diagram picture using MTCNN neural network model and include:
Step 2-1, candidate forms and boundary regressor are obtained using P-Net network, while candidate forms are carried out according to bounding box
Calibration recycles NMS method removal overlapping forms；
Step 2-2, the picture comprising candidate forms for determining P-Net network training in R-Net network, using bounding box to
Amount finely tunes candidate framework, recycles NMS method removal overlapping forms；
Step 2-3, candidate forms are being removed using O-Net network, while is showing five face key point positioning.
3. the face identification method according to claim 2 based on depth convolutional neural networks, it is characterised in that: described
The training of MTCNN neural network model includes three parts, face and non-face classification, and bounding box returns and face position point
Positioning, in which:
Face and non-face classified use cross entropy loss function determine:
；
Wherein,Indicate face probability,Indicate the true tag of background,Indicate the probability of face " prediction " and " the fact is that
Face " degree of closeness,Be worth it is smaller indicate closer, the training objective of the part is to obtain minimum value min()；
Bounding box, which returns to calculate by Euclidean distance, returns loss:
；
Wherein,For the background coordination obtained by neural network forecast,For actual true background coordination,Indicate left
The four-tuple that upper angle, the upper right corner, length and width form,Indicate frame return Euclidean distance, be worth it is smaller, indicate predicted value with
True value is closer, and the training objective of the part is to obtain minimum value min()；
Face position point location minimizes the distance by calculating the Euclidean distance of neural network forecast coordinate and actual coordinate,
Formula are as follows:
；
Wherein,For the face position coordinate obtained by neural network forecast,For actual true face position
Coordinate is set,Indicate ten tuples of five points composition of face,Indicate that the Euclidean distance that frame returns, value are got over
It is small, indicate that predicted value and true value are closer, the training objective of the part is to obtain minimum value min()；
Above-mentioned three parts are integrated to obtain formula:
；
Wherein: N is the quantity of training program,The weight for indicating different task, in P-Net network and R-Net network,=
1,=0.5,=0.5, in O-Net network,=1,=1,=0.5；Indicate sample class
The true tag of type,,Indicate loss function.
4. the face identification method according to claim 1 based on depth convolutional neural networks, it is characterised in that: the step
It is rapid 3) in be inscribed it is oval after, recycle each pixel in the positive face face block diagram picture of face, and judge the pixel whether in ellipse
It is interior, retain if in ellipse, if not ignoring, to obtain oval facial image.
5. the face identification method according to claim 1 based on depth convolutional neural networks, it is characterised in that: the step
It is rapid 5) in human face data information in face template library shift to an earlier date typing, and be also inscribed when typing and oval carry out image in ellipse and mention
It takes.
6. the face identification method according to claim 1 based on depth convolutional neural networks, it is characterised in that: the step
It is rapid 5) in, the eigenmatrix of facial image in face characteristic matrix and face template library is done into cosine similarity operation, obtains original
The similarity value of image in beginning image and face template library.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910017928.3A CN109711384A (en) | 2019-01-09 | 2019-01-09 | A kind of face identification method based on depth convolutional neural networks |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910017928.3A CN109711384A (en) | 2019-01-09 | 2019-01-09 | A kind of face identification method based on depth convolutional neural networks |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109711384A true CN109711384A (en) | 2019-05-03 |
Family
ID=66260989
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910017928.3A Pending CN109711384A (en) | 2019-01-09 | 2019-01-09 | A kind of face identification method based on depth convolutional neural networks |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109711384A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110110673A (en) * | 2019-05-10 | 2019-08-09 | 杭州电子科技大学 | A kind of face identification method based on two-way 2DPCA and cascade feedforward neural network |
| CN110210393A (en) * | 2019-05-31 | 2019-09-06 | 百度在线网络技术（北京）有限公司 | The detection method and device of facial image |
| CN110263731A (en) * | 2019-06-24 | 2019-09-20 | 电子科技大学 | A kind of single step face detection system |
| CN110363091A (en) * | 2019-06-18 | 2019-10-22 | 广州杰赛科技股份有限公司 | Face identification method, device, equipment and storage medium in the case of side face |
| CN110569756A (en) * | 2019-08-26 | 2019-12-13 | 长沙理工大学 | face recognition model construction method, recognition method, device and storage medium |
| CN110728807A (en) * | 2019-09-27 | 2020-01-24 | 深圳市大拿科技有限公司 | Anti-dismantling method of intelligent doorbell and related product |
| CN110738607A (en) * | 2019-09-09 | 2020-01-31 | 平安国际智慧城市科技股份有限公司 | Method, device and equipment for shooting driving license based on artificial intelligence and storage medium |
| CN113283353A (en) * | 2021-05-31 | 2021-08-20 | 创芯国际生物科技(广州)有限公司 | Organoid cell counting method and system based on microscopic image |
| CN113742421A (en) * | 2021-08-20 | 2021-12-03 | 郑州云智信安安全技术有限公司 | Network identity authentication method based on distributed storage and image processing |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105260703A (en) * | 2015-09-15 | 2016-01-20 | 西安邦威电子科技有限公司 | Detection method suitable for smoking behavior of driver under multiple postures |
| SG10201709945UA (en) * | 2016-11-30 | 2018-06-28 | Altumview Systems Inc | Face detection using small-scale convolutional neural network (cnn) modules for embedded systems |
| CN108549886A (en) * | 2018-06-29 | 2018-09-18 | 汉王科技股份有限公司 | A kind of human face in-vivo detection method and device |
| CN108564052A (en) * | 2018-04-24 | 2018-09-21 | 南京邮电大学 | Multi-cam dynamic human face recognition system based on MTCNN and method |
| CN108875602A (en) * | 2018-05-31 | 2018-11-23 | 珠海亿智电子科技有限公司 | Monitor the face identification method based on deep learning under environment |
-
2019
- 2019-01-09 CN CN201910017928.3A patent/CN109711384A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105260703A (en) * | 2015-09-15 | 2016-01-20 | 西安邦威电子科技有限公司 | Detection method suitable for smoking behavior of driver under multiple postures |
| SG10201709945UA (en) * | 2016-11-30 | 2018-06-28 | Altumview Systems Inc | Face detection using small-scale convolutional neural network (cnn) modules for embedded systems |
| CN108564052A (en) * | 2018-04-24 | 2018-09-21 | 南京邮电大学 | Multi-cam dynamic human face recognition system based on MTCNN and method |
| CN108875602A (en) * | 2018-05-31 | 2018-11-23 | 珠海亿智电子科技有限公司 | Monitor the face identification method based on deep learning under environment |
| CN108549886A (en) * | 2018-06-29 | 2018-09-18 | 汉王科技股份有限公司 | A kind of human face in-vivo detection method and device |
Non-Patent Citations (1)
| Title |
|---|
| 陈家骏等: "中学几何词典", 中国人民公安大学出版社, pages: 510 - 511 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110110673A (en) * | 2019-05-10 | 2019-08-09 | 杭州电子科技大学 | A kind of face identification method based on two-way 2DPCA and cascade feedforward neural network |
| CN110210393A (en) * | 2019-05-31 | 2019-09-06 | 百度在线网络技术（北京）有限公司 | The detection method and device of facial image |
| CN110363091A (en) * | 2019-06-18 | 2019-10-22 | 广州杰赛科技股份有限公司 | Face identification method, device, equipment and storage medium in the case of side face |
| CN110363091B (en) * | 2019-06-18 | 2021-08-10 | 广州杰赛科技股份有限公司 | Face recognition method, device and equipment under side face condition and storage medium |
| CN110263731A (en) * | 2019-06-24 | 2019-09-20 | 电子科技大学 | A kind of single step face detection system |
| CN110569756A (en) * | 2019-08-26 | 2019-12-13 | 长沙理工大学 | face recognition model construction method, recognition method, device and storage medium |
| CN110569756B (en) * | 2019-08-26 | 2022-03-22 | 长沙理工大学 | Face recognition model construction method, recognition method, device and storage medium |
| CN110738607A (en) * | 2019-09-09 | 2020-01-31 | 平安国际智慧城市科技股份有限公司 | Method, device and equipment for shooting driving license based on artificial intelligence and storage medium |
| CN110728807A (en) * | 2019-09-27 | 2020-01-24 | 深圳市大拿科技有限公司 | Anti-dismantling method of intelligent doorbell and related product |
| CN113283353A (en) * | 2021-05-31 | 2021-08-20 | 创芯国际生物科技(广州)有限公司 | Organoid cell counting method and system based on microscopic image |
| CN113742421A (en) * | 2021-08-20 | 2021-12-03 | 郑州云智信安安全技术有限公司 | Network identity authentication method based on distributed storage and image processing |
| CN113742421B (en) * | 2021-08-20 | 2023-09-12 | 郑州云智信安安全技术有限公司 | Network identity authentication method based on distributed storage and image processing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109711384A (en) | A kind of face identification method based on depth convolutional neural networks | |
| Chiang et al. | A novel method for detecting lips, eyes and faces in real time | |
| Ban et al. | Face detection based on skin color likelihood | |
| US20210312166A1 (en) | System and method for face recognition based on dynamic updating of facial features | |
| WO2021051601A1 (en) | Method and system for selecting detection box using mask r-cnn, and electronic device and storage medium | |
| CN102194108B (en) | Smile face expression recognition method based on clustering linear discriminant analysis of feature selection | |
| US20050226509A1 (en) | Efficient classification of three dimensional face models for human identification and other applications | |
| CN102902986A (en) | Automatic gender identification system and method | |
| CN104484645B (en) | A kind of " 1 " gesture identification method and system towards man-machine interaction | |
| CN101493887A (en) | Eyebrow image segmentation method based on semi-supervision learning and Hash index | |
| CN108537143B (en) | A kind of face identification method and system based on key area aspect ratio pair | |
| CN112101208A (en) | Feature series fusion gesture recognition method and device for elderly people | |
| WO2021184718A1 (en) | Card border recognition method, apparatus and device, and computer storage medium | |
| CN109558855B (en) | A kind of space gesture recognition methods combined based on palm contour feature with stencil matching method | |
| WO2022160736A1 (en) | Image annotation method and apparatus, electronic device, storage medium and program | |
| JPH04101280A (en) | Face picture collating device | |
| CN109948489A (en) | A kind of face identification system and method based on the fusion of video multiframe face characteristic | |
| CN106980845B (en) | Face key point positioning method based on structured modeling | |
| CN105550642A (en) | Gender identification method and system based on multi-scale linear difference characteristic low-rank expression | |
| CN110222660B (en) | Signature authentication method and system based on dynamic and static feature fusion | |
| CN107357834A (en) | A kind of image search method of view-based access control model conspicuousness fusion | |
| CN111160292A (en) | Human eye detection method | |
| Curran et al. | The use of neural networks in real-time face detection | |
| WO2021151359A1 (en) | Palm print image recognition method, apparatus and device, and computer readable storage medium | |
| CN114445691A (en) | Model training method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190503 |