CN108171223A - A kind of face identification method and system based on multi-model multichannel - Google Patents
A kind of face identification method and system based on multi-model multichannel Download PDFInfo
- Publication number
- CN108171223A CN108171223A CN201810162040.4A CN201810162040A CN108171223A CN 108171223 A CN108171223 A CN 108171223A CN 201810162040 A CN201810162040 A CN 201810162040A CN 108171223 A CN108171223 A CN 108171223A
- Authority
- CN
- China
- Prior art keywords
- face
- feature
- personnel
- image
- identified
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
Abstract
The present invention is a kind of face identification method and system based on multi-model multichannel, it specifically refers to generate different face characteristics using different human face recognition models, different features carries different information, discrimination can largely be improved to carry out recognition of face by merging these features, be related to technical field of face recognition.The present invention obtains facial image by common camera, detects whether face occur using Face datection algorithm, human face region is divided on the basis of having face, the image split is pre-processed.Then feature corresponding to pretreated image zooming-out difference model carries out characteristic processing to these features, personnel to be identified and the similarity of the feature of registered personnel in database are finally measured using COS distance.It is not high that the invention overcomes precision of method of the existing technology, for face environmental change (illumination, posture, is blocked at expression) robustness it is low the shortcomings that, can effectively improve the accuracy rate of recognition of face.
Description
Technical field
The present invention relates to technical field of face recognition, specifically a kind of face identification method based on multi-model multichannel and
System.
Background technology
With the fast development of computer technology and artificial intelligence and growing, the traditional identity of people's demand
Authentication techniques such as fingerprint recognition, iris recognition and speech recognition cannot meet the needs of people.With traditional identity certification skill
Art is compared, and recognition of face has high reliability and the advantage of convenience.
One reliable face identification system is mainly acquired by face, Face datection, pretreatment, feature extraction, classification
Identify this five module compositions.Wherein, feature extraction is most critical, a most important step within the system.However, using single
The face characteristic robustness that extracts of existing feature extraction algorithm it is low, for variation (illumination, expression, the appearance of face environment
State is blocked) it is very sensitive.It can be seen that it is restricted to solve single face characteristic extraction algorithm in the precision of recognition of face
The problem of, it needs to merge various features extraction algorithm at present to improve the accuracy of recognition of face.
Invention content
A kind of the shortcomings that present invention mainly overcomes the above-mentioned prior art, it is proposed that face for having merged a variety of face characteristics
Recognition methods and system.Different models can extract the different features of face, merge a variety of face characteristics and be identified
Certification can improve the precision of recognition of face.
The present invention is that used technical solution is the shortcomings that overcoming conventional method:It provides a kind of based on multi-model multichannel
Face identification method, include the following steps:
Step 1:Facial image is acquired, face image data is obtained using camera, utilizes Viola-Jones Face datections
Whether occur face in algorithm detection image, in the picture there are face on the basis of, face is split, after singulation
It is pre-processed on facial image;
Step 2:On the basis of facial image I after the pre-treatment, it is special to generate AE using the good automatic codec of pre-training
Sign;
Step:3:On the basis of facial image I after the pre-treatment, left eye, right eye, nose are gone out according to the feature points segmentation of face
Son and face image, to each image block extraction HOG, (Histogram of Oriented Gradient, direction gradient are straight
Side's figure) feature.
Step 4:On the basis of facial image I after the pre-treatment, facial image is divided into the image block of several N × N, it is right
Each image block extraction DCP (Dual-Cross Patterns, diesis pattern) feature.
Step:5:AE features, HOG features and DCP feature vectors are merged and utilize WPCA (weighting principal components point
Analysis) reduce redundancy obtain final feature;
Step 6:The cosine between the feature of personnel to be identified and the feature of registered personnel in database is calculated respectively
Distance determines the identity of personnel to be identified according to the size of COS distance.
In step 1, pretreatment stage includes face alignment algorithm (ERT algorithms) and medium filtering figure based on regression tree
As denoising operates.
In step 2, the good autocoder extraction AE features of pre-training, we can be by the pictures of certain in training set
It is expressed as vector x ∈ Rmx1, wherein m is the total pixel number of face picture.The encoder matrix of sparse automatic codec is We=
Rk×m,
Automatically the non-uniform encoding function of codec isSparse coding vector is h
=f (x;We)∈Rk×1, the decoding matrix of automatic codec is Wd∈Rm×k, following unconstrained optimization problem can be established:
Wherein xiRepresent the i-th secondary picture, h in training setiRepresent the sparse coding vector of the i-th secondary picture, λ is regularization system
Number, N are the sum of picture in training set.
For the optimization problem, we using gradient descent method optimize the problem, in this way, can obtain it is trained from
Encoder.
In steps of 5, AE features, HOG features and DCP feature vectors head and the tail are spliced (such as:Vectorial (1,2,3) with to
The spliced result of amount (4,5,6) head and the tail is (1,2,3,4,5,6)) be fused to new feature vector after, WPCA is utilized to reduce number
According to redundancy.
In step 6, it calculates respectively between the feature of personnel to be identified and the feature of registered personnel in database
COS distance, the characteristic vector of chosen distance minimum is recognition result.
The present invention provides a kind of face identification system based on multi-model multichannel, including:
Image collection module, for acquiring facial image to be identified;
Face detection module, for detecting face from the image of acquisition and extracting;
Preprocessing module, for being pre-processed to the face part extracted;
Characteristic extracting module extracts different features using three kinds of models (autocoder, HOG algorithms, DCP algorithms);
Feature processing block splices obtained feature, is normalized and WPCA de-redundancy;
Identification module, for calculating the COS distance between the feature of personnel to be identified and the feature of all registered personnel
And determine the identity information of personnel to be identified.
First, facial image face detection module to be identified is acquired by image collection module, then, from the figure of acquisition
Face is detected by face detection module as in and is extracted, and preprocessing module is passed through to the face part extracted
It is pre-processed;Using characteristic extracting module, the feature for making three kinds of model extractions different;Then, in feature processing block to obtaining
To feature spliced, normalized and WPCA de-redundancy;Later, identification module is used to calculate feature and the institute of personnel to be identified
There is the COS distance between the feature of registered personnel and determine the identity information of personnel to be identified.
The operation that the preprocessing module is related to includes image denoising and face is aligned.
The characteristic extracting module uses three kinds of model extraction face characteristics.
The identification module is calculated in personnel to be identified and database using COS distance between the feature of registered personnel
Distance, identity of the registered personnel identity as personnel to be identified corresponding to the feature of chosen distance minimum.
Advantages of the present invention and effect are:The face identification method and system globe area based on multi-model multichannel provided
The advantages of multiple models, the Fusion Features that different models are generated reach more accurate with reference to the advantage of each model
The purpose of recognition of face.By the test of many face databases, good effect is all achieved, can realize high speed, height
Accurate recognition of face has important meaning in monitoring, security protection etc..
Description of the drawings
Fig. 1 is the flow chart of the present invention.
Fig. 2 is the block diagram of the present invention.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, complete
Site preparation describes, these embodiments are only illustrative of the invention and is not intended to limit the scope of the invention.Based on the reality in the present invention
Apply example, others skilled in the art's all other embodiments obtained without making creative work all belong to
In the scope of protection of the invention.
The embodiment of the present invention proposes a kind of face identification method based on multi-model multichannel, as shown in Figure 1, this method
Including:
(1) clearly human face data is acquired using common camera.
(2) collected facial image is used and whether there is people in Viola-Jones Face datection algorithm detection images
Face, and on the basis of having face, the human face region segmentation that algorithm is detected extracts the image as subsequent experimental.
(3) facial image after being extracted to segmentation carries out pretreatment operation, specifically first will using ERT faces alignment algorithm
Facial image is aligned, and then removes picture noise using medium filtering Image denoising algorithm.
(4) pretreated image is learnt AE features using the good automatic codec of pre-training, i.e., it directly will figure
As input, the coding output of encoder is facial image AE features.The process of the automatic codec models of pre-training is such as
Under:
(5) pictures of certain in training set can be expressed as vector x ∈ R by wemx1, wherein m is the total picture of face picture
Vegetarian refreshments number.The encoder matrix of sparse automatic codec is We=Rk×m, the non-uniform encoding function of automatic codec isSparse coding vector is h=f (x;We)∈Rk×1, the decoding matrix of automatic codec is
Wd∈Rm×k, following unconstrained optimization problem can be established:
Wherein xiRepresent the i-th secondary picture, h in training setiRepresent the sparse coding vector of the i-th secondary picture, λ is regularization system
Number, N are the sum of picture in training set.For the optimization problem we using gradient descent method come be iterated optimization, so
It can obtain trained self-encoding encoder.
(6) to pretreated image, according to the geography information of 68 human face characteristic points by facial image right eye, a left side
Eye, nose and face partial segmentation come out, and extract HOG features respectively to each section, finally that obtain 4 HOG features are first
Tail be stitched together (such as:Vectorial (1,2,3) are (1,2,3,4,5,6) with the spliced result of vectorial (4,5,6) head and the tail) make
HOG features for the face.
(7) to pretreated image, the image block that image propoerly partitioning is several N × N extracts each image block
DCP (Dual-Cross Patterns) feature, several DCP merging features most obtained at last act the DCP spies for being used as the people's face
Sign.
(8) obtained AE features, HOG features and DCP features are stitched together from beginning to end (such as:Vectorial (1,2,3) and vector
(4,5,6) the spliced result of head and the tail is (1,2,3,4,5,6)), and normalize.
(9) characteristic value after normalization is calculated using WPCA (Whitened Principal Compoent Analysis)
Method removes the information redundancy of feature, and formula is:Y=(U Λ-1/2)TX wherein U be by WPCA from training data learning to
Orthogonal intersection cast shadow matrix, Λ=diag { λ1,λ2... for diagonal matrix, λiIth feature value for matrix U.WPCA will be passed through to calculate
Method treated result is as final face feature vector.
(10) calculate respectively personnel to be identified face feature vector and database in registered personnel's face feature vector
Between COS distance, the body of the identity of the corresponding registered personnel of feature vector of chosen distance minimum as personnel to be identified
Part.
Second embodiment of the present invention is related to a kind of face identification system based on multi-model multichannel, as shown in Fig. 2,
Including image collection module, for acquiring facial image to be identified;Face detection module, for being detected from the image of acquisition
Go out face and extract;Preprocessing module, for being pre-processed to the face part extracted;Characteristic extracting module,
Use the different feature of three kinds of model extractions;Feature processing block splices obtained feature, is normalized and WPCA de-redundants
It is remaining;Identification module, for calculating the COS distance between the feature of personnel to be identified and the feature of all registered personnel and true
The identity information of fixed personnel to be identified.
The preprocessing module includes image denoising and face is aligned.
The characteristic extracting module uses multiple model extraction face characteristics, and the feature processing block is to obtained feature
It does and splices, normalizes and WPCA de-redundancy.
The identification module is calculated in personnel to be identified and database using COS distance between the feature of registered personnel
Distance, identity of the registered personnel identity as personnel to be identified corresponding to the feature of chosen distance minimum.
A kind of face identification method and system based on multi-model multichannel provided by the present invention have been carried out in detail above
Thin to introduce, specific case used herein is expounded the principle of the present invention and embodiment, and above example is said
The bright method and its core concept for being merely used to help understand the present invention.
Claims (9)
1. a kind of face identification method based on multi-model multichannel, which is characterized in that include the following steps:
Step 1:Facial image is acquired, face image data is obtained using camera, utilizes Viola-Jones Face datection algorithms
Whether occur face in detection image, in the picture there are face on the basis of, face is split, face after singulation
It is pre-processed on image;
Step 2:On the basis of facial image I after the pre-treatment, AE features are generated using the good automatic codec of pre-training;
Step:3:On the basis of facial image I after the pre-treatment, according to the feature points segmentation of face go out left eye, right eye, nose and
Face image extracts HOG features to each image block;HOG is histograms of oriented gradients;
Step 4:On the basis of facial image I after the pre-treatment, facial image is divided into the image block of several N × N, to each
Image block extracts DCP features;DCP is diesis pattern;
Step:5:AE features, HOG features and DCP feature vectors are merged and weighed PCA WPCA is utilized to reduce
Redundancy obtains final feature;
Step 6:The COS distance between the feature of personnel to be identified and the feature of registered personnel in database is calculated respectively,
The identity of personnel to be identified is determined according to the size of COS distance.
2. a kind of face identification method based on multi-model multichannel according to claim 1, it is characterised in that:In step
In 1, pretreatment stage includes face alignment algorithm and the operation of medium filtering image denoising based on regression tree.
3. a kind of face identification method based on multi-model multichannel according to claim 1, it is characterised in that:In step
In 2, certain pictures in training set are expressed as vector x ∈ R by the good autocoder extraction AE features of pre-trainingmx1;Wherein,
M is the total pixel number of face picture;The encoder matrix of sparse automatic codec is We=Rk×m, automatic codec it is non-
Uniform enconding function isSparse coding vector is h=f (x;We)∈Rk×1, automatic encoding and decoding
The decoding matrix of device is Wd∈Rm×k, establish following unconstrained optimization problem:
Wherein, xiRepresent the i-th secondary picture, h in training setiRepresent the sparse coding vector of the i-th secondary picture, λ is regularization coefficient, N
Sum for picture in training set;
The problem is optimized using gradient descent method for the optimization problem, in this way, obtaining trained self-encoding encoder.
4. a kind of face identification method based on multi-model multichannel according to claim 1, it is characterised in that:In step
In 5, AE features, HOG features and DCP feature vectors head and the tail are spliced, after being fused to new feature vector, number is reduced using WPCA
According to redundancy.
5. a kind of face identification method based on multi-model multichannel according to claim 1, it is characterised in that:In step
In 6, calculate the COS distance between the feature of personnel to be identified and the feature of registered personnel in database respectively, select away from
It is recognition result from minimum characteristic vector.
6. a kind of face identification system based on multi-model multichannel, which is characterized in that including:
Image collection module, for acquiring facial image to be identified;
Face detection module, for detecting face from the image of acquisition and extracting;
Preprocessing module, for being pre-processed to the face part extracted;
Characteristic extracting module extracts different features using autocoder, HOG algorithms, DCP algorithms;
Feature processing block splices obtained feature, is normalized and WPCA de-redundancy;
Identification module, for calculating the COS distance between the feature of personnel to be identified and the feature of all registered personnel and true
The identity information of fixed personnel to be identified;
First, facial image face detection module to be identified is acquired by image collection module, then, from the image of acquisition
Face is detected by face detection module and is extracted, and the face part extracted is carried out by preprocessing module
Pretreatment;Using characteristic extracting module, the feature for making three kinds of model extractions different;Then, in feature processing block to obtaining
Feature spliced, is normalized and WPCA de-redundancy;Later, identification module be used to calculating the feature of personnel to be identified with it is all
The identity information of COS distance and determining personnel to be identified between the feature of accredited personnel.
7. a kind of face identification system based on multi-model multichannel according to claim 6, it is characterised in that:It is described pre-
The operation that processing module is related to includes image denoising and face is aligned.
8. a kind of face identification system based on multi-model multichannel according to claim 6, it is characterised in that:The knowledge
Other module calculates the distance between personnel to be identified and the feature of registered personnel in database, chosen distance using COS distance
Identity of the registered personnel identity as personnel to be identified corresponding to minimum feature.
9. a kind of face identification system based on multi-model multichannel according to claim 6, it is characterised in that:Normalization
Characteristic value afterwards removes the information redundancy of feature using WPCA algorithms, and formula is:Y=(U Λ-1/2)Tx;Wherein, U is to pass through
WPCA from training data learning to orthogonal intersection cast shadow matrix, Λ=diag { λ1,λ2... for diagonal matrix, λiFor matrix U
Ith feature value;Using the result after WPCA algorithm process as final face feature vector.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810162040.4A CN108171223A (en) | 2018-02-27 | 2018-02-27 | A kind of face identification method and system based on multi-model multichannel |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810162040.4A CN108171223A (en) | 2018-02-27 | 2018-02-27 | A kind of face identification method and system based on multi-model multichannel |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108171223A true CN108171223A (en) | 2018-06-15 |
Family
ID=62511653
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810162040.4A Pending CN108171223A (en) | 2018-02-27 | 2018-02-27 | A kind of face identification method and system based on multi-model multichannel |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108171223A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109190535A (en) * | 2018-08-23 | 2019-01-11 | 南京邮电大学 | A kind of face blee analysis method and system based on deep learning |
CN109376602A (en) * | 2018-09-21 | 2019-02-22 | 厦门中控智慧信息技术有限公司 | A kind of finger vein identification method, device and terminal device |
CN109398310A (en) * | 2018-09-26 | 2019-03-01 | 深圳万智联合科技有限公司 | A kind of pilotless automobile |
CN109472240A (en) * | 2018-11-12 | 2019-03-15 | 北京影谱科技股份有限公司 | Recognition of face multi-model self-adapting Fusion Features Enhancement Method and device |
CN110427265A (en) * | 2019-07-03 | 2019-11-08 | 平安科技(深圳)有限公司 | Method, apparatus, computer equipment and the storage medium of recognition of face |
CN110941991A (en) * | 2019-10-28 | 2020-03-31 | 成都华迈通信技术有限公司 | Multi-channel structured data acquisition system and data acquisition method |
CN111724522A (en) * | 2020-05-25 | 2020-09-29 | 浙江大华技术股份有限公司 | Access control system, method and device, control equipment and storage medium |
WO2021057426A1 (en) * | 2019-09-29 | 2021-04-01 | 腾讯科技(深圳)有限公司 | Method and apparatus for training image fusion processing model, device, and storage medium |
CN115019373A (en) * | 2022-06-30 | 2022-09-06 | 北京瑞莱智慧科技有限公司 | Method, device and storage medium for tracking and detecting specific person |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105718868A (en) * | 2016-01-18 | 2016-06-29 | 中国科学院计算技术研究所 | Face detection system and method for multi-pose faces |
CN105844132A (en) * | 2016-03-17 | 2016-08-10 | 中国科学院上海微系统与信息技术研究所 | Mobile terminal-based human face identification method and system |
CN105913025A (en) * | 2016-04-12 | 2016-08-31 | 湖北工业大学 | Deep learning face identification method based on multiple-characteristic fusion |
CN106169072A (en) * | 2016-07-07 | 2016-11-30 | 中国科学院上海微系统与信息技术研究所 | A kind of face identification method based on Taylor expansion and system |
CN106203354A (en) * | 2016-07-14 | 2016-12-07 | 南京信息工程大学 | Scene recognition method based on interacting depth structure |
CN106874879A (en) * | 2017-02-21 | 2017-06-20 | 华南师范大学 | Handwritten Digit Recognition method based on multiple features fusion and deep learning network extraction |
CN107528832A (en) * | 2017-08-04 | 2017-12-29 | 北京中晟信达科技有限公司 | Baseline structure and the unknown anomaly detection method of a kind of system-oriented daily record |
CN107622225A (en) * | 2017-07-27 | 2018-01-23 | 成都信息工程大学 | Face identification method based on independent component analysis network |
-
2018
- 2018-02-27 CN CN201810162040.4A patent/CN108171223A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105718868A (en) * | 2016-01-18 | 2016-06-29 | 中国科学院计算技术研究所 | Face detection system and method for multi-pose faces |
CN105844132A (en) * | 2016-03-17 | 2016-08-10 | 中国科学院上海微系统与信息技术研究所 | Mobile terminal-based human face identification method and system |
CN105913025A (en) * | 2016-04-12 | 2016-08-31 | 湖北工业大学 | Deep learning face identification method based on multiple-characteristic fusion |
CN106169072A (en) * | 2016-07-07 | 2016-11-30 | 中国科学院上海微系统与信息技术研究所 | A kind of face identification method based on Taylor expansion and system |
CN106203354A (en) * | 2016-07-14 | 2016-12-07 | 南京信息工程大学 | Scene recognition method based on interacting depth structure |
CN106874879A (en) * | 2017-02-21 | 2017-06-20 | 华南师范大学 | Handwritten Digit Recognition method based on multiple features fusion and deep learning network extraction |
CN107622225A (en) * | 2017-07-27 | 2018-01-23 | 成都信息工程大学 | Face identification method based on independent component analysis network |
CN107528832A (en) * | 2017-08-04 | 2017-12-29 | 北京中晟信达科技有限公司 | Baseline structure and the unknown anomaly detection method of a kind of system-oriented daily record |
Non-Patent Citations (5)
Title |
---|
乔宇等: "基于加权主元分析(WPCA)的人脸识别", 《重庆大学学报(自然科学版)》 * |
刘海媚等: "基于方向边缘幅值的尺度块 LBP 人脸识别", 《计算机工程与应用》 * |
段龙云等: "多任务姿态不变的人脸识别", 《电脑知识与技术》 * |
肖冬杰等: "基于稀疏表示的自适应图像融合方法研究", 《计算机应用与软件》 * |
陈明亮等: "彩色自然场景下的多特征融合车辆检测算法", 《系统仿真技术》 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109190535A (en) * | 2018-08-23 | 2019-01-11 | 南京邮电大学 | A kind of face blee analysis method and system based on deep learning |
CN109190535B (en) * | 2018-08-23 | 2022-04-01 | 南京邮电大学 | Face complexion analysis method and system based on deep learning |
CN109376602A (en) * | 2018-09-21 | 2019-02-22 | 厦门中控智慧信息技术有限公司 | A kind of finger vein identification method, device and terminal device |
CN109398310A (en) * | 2018-09-26 | 2019-03-01 | 深圳万智联合科技有限公司 | A kind of pilotless automobile |
CN109472240A (en) * | 2018-11-12 | 2019-03-15 | 北京影谱科技股份有限公司 | Recognition of face multi-model self-adapting Fusion Features Enhancement Method and device |
CN110427265A (en) * | 2019-07-03 | 2019-11-08 | 平安科技(深圳)有限公司 | Method, apparatus, computer equipment and the storage medium of recognition of face |
WO2021057426A1 (en) * | 2019-09-29 | 2021-04-01 | 腾讯科技(深圳)有限公司 | Method and apparatus for training image fusion processing model, device, and storage medium |
US11526712B2 (en) | 2019-09-29 | 2022-12-13 | Tencent Technology (Shenzhen) Company Limited | Training method and apparatus for image fusion processing model, device, and storage medium |
CN110941991A (en) * | 2019-10-28 | 2020-03-31 | 成都华迈通信技术有限公司 | Multi-channel structured data acquisition system and data acquisition method |
CN110941991B (en) * | 2019-10-28 | 2023-05-23 | 成都华迈通信技术有限公司 | Multichannel structured data acquisition system and data acquisition method |
CN111724522A (en) * | 2020-05-25 | 2020-09-29 | 浙江大华技术股份有限公司 | Access control system, method and device, control equipment and storage medium |
CN115019373A (en) * | 2022-06-30 | 2022-09-06 | 北京瑞莱智慧科技有限公司 | Method, device and storage medium for tracking and detecting specific person |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108171223A (en) | A kind of face identification method and system based on multi-model multichannel | |
William et al. | Face recognition using facenet (survey, performance test, and comparison) | |
CN109558810B (en) | Target person identification method based on part segmentation and fusion | |
US6661907B2 (en) | Face detection in digital images | |
Kumar et al. | Real time face recognition using adaboost improved fast PCA algorithm | |
CN108563999A (en) | A kind of piece identity's recognition methods and device towards low quality video image | |
CN111126240B (en) | Three-channel feature fusion face recognition method | |
Anand et al. | An improved local binary patterns histograms techniques for face recognition for real time application | |
CN110796101A (en) | Face recognition method and system of embedded platform | |
CN106650574A (en) | Face identification method based on PCANet | |
CN107038400A (en) | Face identification device and method and utilize its target person tracks of device and method | |
Shrivastava et al. | Conceptual model for proficient automated attendance system based on face recognition and gender classification using Haar-Cascade, LBPH algorithm along with LDA model | |
Barbu | An automatic face detection system for RGB images | |
Diyasa et al. | Multi-face Recognition for the Detection of Prisoners in Jail using a Modified Cascade Classifier and CNN | |
Galiyawala et al. | Person retrieval in surveillance videos using deep soft biometrics | |
Hrkać et al. | Deep learning architectures for tattoo detection and de-identification | |
Mohamed et al. | Automated face recogntion system: Multi-input databases | |
Gürel | Development of a face recognition system | |
Das et al. | Human face detection in color images using HSV color histogram and WLD | |
Jacob | Comparison of popular face detection and recognition techniques | |
KR101174103B1 (en) | A face recognition method of Mathematics pattern analysis for muscloskeletal in basics | |
Curran et al. | The use of neural networks in real-time face detection | |
CN111428670B (en) | Face detection method, face detection device, storage medium and equipment | |
Gupta et al. | HaarCascade and LBPH Algorithms in Face Recognition Analysis | |
Sivapriyan et al. | Analysis of facial recognition techniques |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180615 |