CN114694008A - Remote face recognition system - Google Patents
Remote face recognition system Download PDFInfo
- Publication number
- CN114694008A CN114694008A CN202111645008.XA CN202111645008A CN114694008A CN 114694008 A CN114694008 A CN 114694008A CN 202111645008 A CN202111645008 A CN 202111645008A CN 114694008 A CN114694008 A CN 114694008A
- Authority
- CN
- China
- Prior art keywords
- module
- face
- image
- data
- face recognition
- 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
- 238000012545 processing Methods 0.000 claims abstract description 33
- 238000012549 training Methods 0.000 claims abstract description 10
- 230000001815 facial effect Effects 0.000 claims description 37
- 239000011159 matrix material Substances 0.000 claims description 26
- 238000004891 communication Methods 0.000 claims description 24
- 238000012544 monitoring process Methods 0.000 claims description 23
- 230000009467 reduction Effects 0.000 claims description 23
- 238000001514 detection method Methods 0.000 claims description 22
- 230000005540 biological transmission Effects 0.000 claims description 16
- 238000011156 evaluation Methods 0.000 claims description 16
- 238000000034 method Methods 0.000 claims description 16
- 238000005516 engineering process Methods 0.000 claims description 14
- 238000000605 extraction Methods 0.000 claims description 13
- 238000007781 pre-processing Methods 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 9
- 239000013598 vector Substances 0.000 claims description 9
- 239000000284 extract Substances 0.000 claims description 4
- 230000001502 supplementing effect Effects 0.000 claims description 4
- 230000000694 effects Effects 0.000 claims description 3
- 238000009432 framing Methods 0.000 claims description 3
- 230000003993 interaction Effects 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000000007 visual effect Effects 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 2
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 230000001131 transforming effect Effects 0.000 claims 1
- 238000011161 development Methods 0.000 abstract description 3
- 210000000887 face Anatomy 0.000 description 15
- 230000006872 improvement Effects 0.000 description 9
- 238000004590 computer program Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000010276 construction Methods 0.000 description 4
- 239000000047 product Substances 0.000 description 3
- 238000000354 decomposition reaction Methods 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000002265 prevention Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- QNRATNLHPGXHMA-XZHTYLCXSA-N (r)-(6-ethoxyquinolin-4-yl)-[(2s,4s,5r)-5-ethyl-1-azabicyclo[2.2.2]octan-2-yl]methanol;hydrochloride Chemical compound Cl.C([C@H]([C@H](C1)CC)C2)CN1[C@@H]2[C@H](O)C1=CC=NC2=CC=C(OCC)C=C21 QNRATNLHPGXHMA-XZHTYLCXSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Collating Specific Patterns (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of face recognition, in particular to a remote face recognition system. The system comprises an infrastructure unit, a data processing unit, a face recognition unit and an application management unit; the infrastructure unit is used for managing and controlling the foundation of the operation of the support system; the data processing unit is used for acquiring and processing an image containing a human face; the face recognition unit is used for recognizing a face; the application management unit is used for managing the functional service of the system according to different application scenes. The design of the invention can realize real-time remote face recognition; the personnel identity can be more accurately identified, and through the label identification and the feedback reminding means, relevant workers can pertinently carry out service work, and the body feeling of the enterprise client is improved; different parameters can be set according to different application scenes and training of different face characteristics can be carried out, the application range of the system is effectively expanded, the development of a smart city can be promoted, and the service management level of government affairs and enterprises is improved.
Description
Technical Field
The invention relates to the technical field of face recognition, in particular to a remote face recognition system.
Background
The face recognition technology belongs to one of biological feature recognition technologies, and particularly relates to a computer technology which utilizes analysis and comparison to judge whether a face exists or not based on the face features of people, further give the position and size of each face and the position information of each main facial organ, extract the identity features of each face according to the information, and compare the identity features with the known faces to achieve the purpose of recognizing the identity of each face. The generalized face recognition comprises a series of related technologies such as face image acquisition, face positioning, face recognition preprocessing, identity confirmation, identity search and the like. At present, the face recognition technology is mainly applied to access control management, electronic product unlocking, electronic payment and the like, the application scenes generally require that a user is close to an identification terminal, the user is required to actively perform the face recognition technology, and the application limit is large. With the rapid development of smart cities, the fields of smart traffic, smart security, smart business service management and the like are continuously developed, if the face recognition technology can be applied to the fields and the identities of related people are remotely recognized, the management level of the smart city is expected to be improved, so that workers of related enterprises can perform service work in a targeted manner, and a large amount of useless work is reduced. However, currently, there is no widely applicable remote face recognition system.
Disclosure of Invention
The present invention is directed to a remote face recognition system to solve the above problems in the background art.
To achieve the above technical problem, one of the objectives of the present invention is to provide a remote face recognition system, which includes
The system comprises an infrastructure unit, a data processing unit, a face recognition unit and an application management unit; the infrastructure unit, the data processing unit, the face recognition unit and the application management unit are sequentially connected through network communication; the infrastructure unit is used for managing, controlling and distributing equipment and data bases for operation of the support system; the data processing unit is used for acquiring images containing a large number of human faces from the image data and processing the images; the face recognition unit is used for recognizing the face in the image through various face recognition technologies to judge the identity of the face; the application management unit is used for managing the functional service of the face recognition system according to different application scenes;
the infrastructure unit comprises an equipment management module, a basic database module, a parameter presetting module and a communication support module;
the data processing unit comprises an image shooting and recording module, a lossless transmission module, an image management module and a distribution storage module;
the face recognition unit comprises a recognition mode module, a comparison recognition module, an evaluation sorting module and a result output module;
the application management unit comprises an application scene module, a label display module, a feedback reminding module and a supplement updating module.
As a further improvement of the technical scheme, the equipment management module, the basic database module and the parameter presetting module are sequentially connected through network communication; the equipment management module is used for performing connection control and management on various kinds of equipment which are constructed into a network architecture supporting system operation; the basic database module is used for establishing a database through a data manager so as to store face images and related information of related personnel which are pre-recorded; the parameter presetting module is used for presetting label information of related personnel information and related limiting parameters in the system operation process; the communication support module is used for building signal connection and data transmission channels among all devices and all layers of the system through various communication means.
The basic equipment comprises a computer, a processor, a display, a monitoring camera and the like.
The data contained in the basic database is derived from face information and related public information of people who register by online publicists, famous people, enterprise clients and enterprise workers.
The preset parameters comprise label data of fixed crowds, the number of faces which can be identified simultaneously in one monitoring camera, the data volume which can be processed by the system in parallel, the time limit of a group of face identification operations and the like.
The communication means includes a wired network, a local area network, a work private network, a wireless network, and the like.
As a further improvement of the technical solution, a signal output end of the image capturing module is connected with a signal input end of the lossless transmission module, a signal output end of the lossless transmission module is connected with a signal input end of the image management module, and a signal output end of the image management module is connected with a signal input end of the distributed storage module; the image shooting module is used for shooting an image video in the visual field range of the monitoring camera in real time; the lossless transmission module is used for transmitting the shot image data to a processing layer of the system in a real-time and lossless manner; the image management module is used for intercepting images containing human faces from the video images and carrying out treatment management on the images so as to facilitate the identification and judgment of human face identities; the distributed storage module is used for respectively storing and managing the original image and various types of image data generated in the processing process.
As a further improvement of the technical scheme, the image management module comprises an image preprocessing module, a PCA dimension reduction module, a feature extraction module, a feature classification module and a cascade detection module; the signal output end of the graphics preprocessing module is connected with the signal input end of the PCA dimension reduction module, the signal output end of the PCA dimension reduction module is connected with the signal input end of the feature extraction module, the signal output end of the feature extraction module is connected with the signal input end of the feature classification module, and the signal output end of the feature classification module is connected with the signal input end of the cascade detection module; the image preprocessing module is used for intercepting a clear image containing a human face from a video image, and performing processing operations such as cutting, scaling, image enhancement, color binarization and the like on the image; the PCA dimension reduction module is used for converting the original data into a group of representations which are linearly independent of each dimension through linear transformation so as to extract main characteristic components of the data and realize the dimension reduction of the high-dimensional data; the feature extraction module is used for collecting a face image of a user and extracting facial feature points in the image; the feature classification module is used for placing the extracted feature points into a linear SVM for classification so as to train an SVM model which can distinguish whether the image is a face image or not; the cascade detection module is used for carrying out face detection and correction operation on the image through a cascade detection classifier.
As a further improvement of the technical solution, in the PCA dimension reduction module, the PCA dimension reduction algorithm includes the following steps:
inputting: training sample set $ D ═ x{(1)}，x{(2)}，...，xΛ{ (m) } }, a low-dimensional spatial dimension d' $;
the algorithm process is as follows:
step1, centralizing all samples, namely, performing a mean value removing operation:
step2, calculating covariance matrix XX of sampleT；
Step3, Pair covariance matrix XXTCarrying out characteristic value decomposition;
step4, and taking the eigenvector w corresponding to the largest d' eigenvalues1，w2，...，wd′；
Step5, multiplying the original matrix by the projection matrix: x · W, i.e. the dimensionality reduced dataset X', where X is mxn-dimensional and W ═ W1，w2，...，wd′]Dimension n x d';
and (3) outputting: the dimensionality reduced dataset X'.
Wherein the expression of the covariance matrix is as follows:
(1) characteristic XiAnd feature XjHas a covariance of：
Wherein the content of the first and second substances,
represents a feature Xi，XjIs taken from the kth sample, and
then it is the sample mean representing both features;
it can be seen that when X isi＝XjWhen the variance is zero, the covariance is the variance;
(2) for a sample with only two features, the covariance matrix is:
when the number of features is n, the covariance matrix is an n × n-dimensional matrix, and the diagonal is the variance value of each feature.
Wherein, the definition of the characteristic vector and the characteristic value is as follows:
for matrix a, if a ξ ═ λ is satisfiedζZeta is the eigenvector of the matrix A, and lambda is the eigenvalue of the matrix A; and sorting the eigenvalues in the descending order, and selecting the eigenvectors corresponding to the first k eigenvalues as the projection vector to be obtained.
As a further improvement of the technical solution, in the cascade detection module, an expression of a simple cascade detection classifier is as follows:
in the formula, ciRepresented by the ith weak classifier, X by the feature vector, F by the classification score, each ciA classification result is output for X according to its own classification method.
As a further improvement of the technical solution, a signal output end of the identification mode module is connected with a signal input end of the comparison identification module, a signal output end of the comparison identification module is connected with a signal input end of the evaluation sorting module, and a signal output end of the evaluation sorting module is connected with a signal input end of the result output module; the recognition mode module is used for calling different algorithms to perform face recognition operations in different modes; the comparison and identification module is used for comparing the processed facial information with a facial image in a user database; the evaluation sorting module is used for calculating the similarity of two faces in the acquired images and the images of the database according to the percentage, and sorting a plurality of pieces of face data with similarity from high to low according to the similarity; and the result output module is used for outputting the face identity result with the recognition similarity reaching hundreds and the corresponding related information.
As a further improvement of the technical solution, the recognition mode module comprises a facial feature module, a facial form and eye form module, a facial picture recognition module and a wearing recognition module; the facial feature module, the facial form and eye shape module, the face recognition module and the wearing recognition module are sequentially connected through network communication and run in parallel; the facial feature module is used for accurately identifying the identity of the face through the facial features by detecting the positioning of each key point 72 such as eye, mouth, nose contour and the like of the face in the image; the face shape and eye shape module is used for performing key point positioning analysis on the eye shape, the eye distance and the face shape of the face after the face is detected so as to obtain the facial feature data of the face shape and eye shape, and therefore the judgment of face recognition can be more accurately performed; the facial recognition module is used for marking and analyzing the face, acquiring key feature points of the face, accurately recognizing various facial features and combining psychological needs of users to be suitable for the drainage of marketing interaction activities; the wearing identification module is used for identifying and training data based on large-scale wearing of safety articles, automatically identifying the wearing condition of the safety articles of field personnel by matching with a field camera and combining with a face identification technology, and corresponding the result to people to realize accurate and efficient supervision.
The types of the facial shapes and the eyes comprise round faces, square faces, long faces, inverted triangular faces, rhombic faces, upper eyes, lower eyes, round eyes and the like.
The object to be worn and identified can be a safety helmet (suitable for large-scale construction places), a mask (suitable for medical institutions, occasions where most people gather in epidemic situations, stations and the like, and also suitable for identification of the person wearing the mask), a safety belt and the like.
As a further improvement of the technical solution, in the evaluation ranking module, a cosine similarity algorithm is adopted when evaluating the similarity between two faces, and the algorithm is defined as:
wherein:
in the formula, data x and y each have n multivariate attributes.
Specifically, from a geometric perspective, the cosine similarity is an included angle between two vectors; and after the included angle of the cosine similarity is obtained, converting the degree of the included angle into percentage.
As a further improvement of the technical solution, the application scene module, the label display module, the feedback reminding module and the supplementary update module are sequentially connected through network communication; the application scene module is used for presetting application scenes of the system for selection, setting different operation parameters aiming at each application scene and training face data with different emphasis characteristics; the label display module is used for framing the identity of the identified person in the monitoring video and carrying out label display on the crowd label to which the person belongs; the feedback reminding module is used for feeding back the identity of a specific crowd to a worker in various ways when the identity of the specific crowd is recognized; the supplementary updating module is used for automatically or manually supplementing and updating the face data and the related information of the new person recognized by the system for the first time into a database of the system.
The types of the application scenes comprise intelligent security, enterprise VIP customer management, unlocking access control, financial and government affair service, identity authentication, public place health and safety management, large-scale engineering site management, intelligent city service, intelligent traffic management and the like.
The crowd labels comprise VIP customers (suitable for business enterprises, building sales departments of local manufacturers and the like), famous people/actors/stars/singers (suitable for order management in public places), credit blacklists (suitable for financial institutions, transportation ticket selling and the like), inertial thieves (suitable for business monitoring, public transportation, security and protection management and the like), unworn safety helmets/masks/safety belts/helmets (suitable for construction sites, epidemic prevention and control, intelligent traffic management and the like) and the like.
The feedback reminding mode comprises a popup window, a warning sound, a short message and the like.
The second objective of the present invention is to provide an operation method of a remote face recognition system, comprising:
the method comprises the steps of firstly, building a network architecture and a system platform which comprise a processor, a database, a monitoring camera and the like, carrying out installation connection and trial operation on each basic device, selecting an application scene of the system, setting relevant limiting parameters according to the application scene, importing or inputting basic data, shooting an image video in real time through the monitoring camera and transmitting the image video to a system processing layer when the system is in operation, extracting an image containing a human face from the video by the processor, preprocessing the image, extracting human face characteristics, comparing the human face characteristics with human face data stored in the database, outputting a recognition result, carrying out mark display on the recognized human face identity or a corresponding crowd label on the video, and simultaneously feeding back the recognition result to relevant managers.
The present invention also provides an operating apparatus of a remote face recognition system, which includes a processor, a memory, and a computer program stored in the memory and running on the processor, wherein the processor is configured to implement any one of the above remote face recognition systems when executing the computer program.
It is a fourth object of the present invention to provide a computer-readable storage medium, which stores a computer program that, when executed by a processor, implements any of the above-described remote face recognition systems.
Compared with the prior art, the invention has the beneficial effects that:
1. the remote face recognition system is provided with a system platform which can be connected with a plurality of monitoring cameras, and a plurality of face recognition algorithms are loaded, so that face information in a monitoring video can be contrasted and recognized in real time, and remote face recognition is realized;
2. the remote face recognition system can more accurately recognize the identity of personnel through face recognition modes in various modes compared with the traditional face recognition technology, and enables related workers to pertinently carry out service work through label identification and feedback reminding means, thereby reducing waste of useless work and improving the body feeling of business clients;
3. this long-range face identification system can set up different parameters and carry out the training of different face characteristics according to the application scene of difference through setting up multiple application scene, has effectively enlarged the range of application of system, can promote the development in wisdom city, improves the service management level of government affairs, business.
Drawings
FIG. 1 is a block diagram of an exemplary overall system product architecture of the present invention;
FIG. 2 is a block diagram of the overall system apparatus of the present invention;
FIG. 3 is a diagram of one embodiment of a local system device architecture;
FIG. 4 is a second block diagram of a local system apparatus according to the present invention;
FIG. 5 is a third block diagram of a local system apparatus according to the present invention;
FIG. 6 is a fourth embodiment of the present invention;
FIG. 7 is a fifth embodiment of the present invention;
FIG. 8 is a sixth embodiment of the present invention;
fig. 9 is a schematic structural diagram of an exemplary cloud platform apparatus according to the present invention.
The various reference numbers in the figures mean:
1. a computing processor; 2. face recognition technology software; 3. a face database; 4. a surveillance camera; 5. displaying a large screen;
100. an infrastructure unit; 101. a device management module; 102. a basic database module; 103. a parameter presetting module; 104. A communication support module;
200. a data processing unit; 201. an image capturing module; 202. a lossless transmission module; 203. an image management module; 2031. A graphics pre-processing module; 2032. a PCA dimension reduction module; 2033. a feature extraction module; 2034. a feature classification module; 2035. a cascade detection module; 204. a distributed storage module;
300. a face recognition unit; 301. a recognition mode module; 3011. a facial feature module; 3012. a facial eye-shaped module; 3013. a face recognition module; 3014. a wear identification module; 302. a comparison identification module; 303. an evaluation sequencing module; 304. a result output module;
400. an application management unit; 401. an application scenario module; 402. a label display module; 403. a feedback reminding module; 404. And supplementing the updating module.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
As shown in fig. 1 to 9, the present embodiment provides a remote face recognition system, which includes
The system comprises an infrastructure unit 100, a data processing unit 200, a face recognition unit 300 and an application management unit 400; the infrastructure unit 100, the data processing unit 200, the face recognition unit 300 and the application management unit 400 are connected in sequence through network communication; the infrastructure unit 100 is configured to manage and control and allocate devices and data bases for supporting system operation; the data processing unit 200 is used for acquiring images containing a large number of human faces from the image data and processing the images; the face recognition unit 300 is configured to recognize a face in an image through a plurality of face recognition technologies to determine an identity of the face; the application management unit 400 is configured to manage functional services of the face recognition system according to different application scenarios;
the infrastructure unit 100 comprises a device management module 101, a basic database module 102, a parameter presetting module 103 and a communication support module 104;
the data processing unit 200 comprises an image shooting module 201, a lossless transmission module 202, an image management module 203 and a distribution storage module 204;
the face recognition unit 300 comprises a recognition mode module 301, a comparison recognition module 302, an evaluation sorting module 303 and a result output module 304;
the application management unit 400 includes an application scenario module 401, a tag presentation module 402, a feedback alert module 403, and a supplementary update module 404.
In this embodiment, the device management module 101, the basic database module 102, and the parameter presetting module 103 are connected in sequence through network communication; the device management module 101 is configured to perform connection control and management on various devices that construct a network architecture supporting system operation; the basic database module 102 is used for establishing a database through a data manager so as to store face images and related information of related personnel which are pre-recorded; the parameter presetting module 103 is used for presetting label information of related personnel information and related limiting parameters in the system operation process; the communication support module 104 is used for establishing signal connection and data transmission channels among various devices and layers of the system through various communication means.
The basic equipment comprises a computer, a processor, a display, a monitoring camera and the like.
The data contained in the basic database is derived from face information and related public information of people who register by online publicists, famous people, enterprise clients and enterprise workers.
The preset parameters comprise label data of fixed crowds, the number of faces which can be identified simultaneously in one monitoring camera, the data volume which can be processed by the system in parallel, the time limit of a group of face identification operations and the like.
The communication means includes a wired network, a local area network, a work private network, a wireless network, and the like.
In this embodiment, the signal output end of the image capturing module 201 is connected to the signal input end of the lossless transmission module 202, the signal output end of the lossless transmission module 202 is connected to the signal input end of the image management module 203, and the signal output end of the image management module 203 is connected to the signal input end of the distributed storage module 204; the image shooting and recording module 201 is used for shooting an image video in the visual field range of the monitoring camera in real time; the lossless transmission module 202 is used for transmitting the shot image data to a processing layer of the system in a real-time and lossless manner; the image management module 203 is configured to intercept an image containing a face from a video image, and perform processing management on the image so as to perform identification and determination of the face identity; the distribution storage module 204 is configured to separately store and manage the original image and various types of image data generated in the processing process.
Further, the image management module 203 includes an image preprocessing module 2031, a PCA dimension reduction module 2032, a feature extraction module 2033, a feature classification module 2034, and a cascade detection module 2035; the signal output end of the graph preprocessing module 2031 is connected to the signal input end of the PCA dimension reduction module 2032, the signal output end of the PCA dimension reduction module 2032 is connected to the signal input end of the feature extraction module 2033, the signal output end of the feature extraction module 2033 is connected to the signal input end of the feature classification module 2034, and the signal output end of the feature classification module 2034 is connected to the signal input end of the cascade detection module 2035; the graphics preprocessing module 2031 is configured to intercept a clear image containing a human face from a video image, and perform processing operations such as cropping, scaling, image enhancement, color binarization, and the like on the image; the PCA dimension reduction module 2032 is configured to transform the raw data into a set of representations linearly independent of each dimension through linear transformation to extract principal feature components of the data to achieve dimension reduction of the high-dimensional data; the feature extraction module 2033 is configured to collect a face image of a user and extract facial feature points in the image; the feature classification module 2034 is configured to place the extracted feature points in a linear SVM for classification so as to train an SVM model that can distinguish whether an image is a face image; the cascade detection module 2035 is used for performing face detection and rectification operations on the image by the cascade detection classifier.
Specifically, in the PCA dimension reduction module 2032, the PCA dimension reduction algorithm includes the following steps:
inputting: training sample set $ D ═ x{(1)}，x{(2)}，...，xΛ{ (m) } }, a low-dimensional spatial dimension d' $;
the algorithm process is as follows:
step1, centralizing all samples, namely, performing a mean value removing operation:
step2, calculating covariance matrix XX of sampleT；
Step3, Pair covariance matrix XXTCarrying out characteristic value decomposition;
step4, and taking the eigenvector w corresponding to the largest d' eigenvalues1，w2，...，wd′；
Step5, multiplying the original matrix by the projection matrix: x · W, i.e. the dimensionality reduced dataset X', where X is mxn-dimensional and W ═ W1，w2，...，wd′]Is dimension n x d';
and (3) outputting: and D, reducing the dimension of the data set X'.
Wherein the expression of the covariance matrix is as follows:
(1) characteristic XiAnd feature XjThe covariance of (a) is:
wherein the content of the first and second substances,
represents a feature Xi，XjIs taken from the kth sample, and
then it is the sample mean representing both features;
it can be seen that when X isi＝XjThen, the covariance is the variance;
(2) for a sample with only two features, the covariance matrix is:
when the number of features is n, the covariance matrix is an n × n-dimensional matrix, and the diagonal is the variance value of each feature.
Wherein, the definition of the characteristic vector and the characteristic value is as follows:
for matrix a, if a ξ ═ λ is satisfiedζZeta is the eigenvector of the matrix A, and lambda is the eigenvalue of the matrix A; and sorting the eigenvalues in the descending order, and selecting the eigenvectors corresponding to the first k eigenvalues as the projection vector to be obtained.
Specifically, in the cascade detection module 2035, the expression of the simple cascade detection classifier is:
in the formula, ciRepresented by the ith weak classifier, X by the feature vector, F by the classification score, each ciA classification result is output for X according to its own classification method.
In this embodiment, the signal output end of the identification mode module 301 is connected to the signal input end of the comparison identification module 302, the signal output end of the comparison identification module 302 is connected to the signal input end of the evaluation sorting module 303, and the signal output end of the evaluation sorting module 303 is connected to the signal input end of the result output module 304; the recognition mode module 301 is configured to call different algorithms to perform face recognition operations in different modes; the comparison and recognition module 302 is configured to compare the processed facial information with a facial image in a user database; the evaluation sorting module 303 is configured to calculate similarity between two faces in the acquired image and the image in the database in percentage, and sort a plurality of pieces of face data having similarity from high to low according to the similarity; the result output module 304 is configured to output a face identity result with a similarity reaching one hundred percent and related information corresponding to the face identity result.
Further, the recognition mode module 301 includes a facial feature module 3011, a face and eye type module 3012, a face recognition module 3013, and a wearing recognition module 3014; the face feature module 3011, the face-type eye-type module 3012, the face recognition module 3013 and the wearing recognition module 3014 are connected in sequence through network communication and operate in parallel; the facial feature module 3011 is configured to determine the identity of a human face accurately through facial features by detecting the location of each key point 72 of the eye, mouth, nose contour, etc. of the human face in an image; the face-type eye-type module 3012 is configured to perform key point positioning analysis on the eye shape, eye distance, and face shape of the face after the face is detected, so as to obtain facial feature data of the face-type eye shape, thereby performing determination of face recognition more accurately; the facial recognition module 3013 is configured to perform labeling analysis on the face, obtain key feature points of the face, accurately recognize various facial features, and apply to drainage of marketing interaction activities in combination with psychological needs of users; the wearing identification module 3014 is used for data identification training based on wearing security articles on a large scale, matching with a field camera and combining with a face identification technology, automatically identifying the wearing condition of the security articles of field personnel, and corresponding the result to people to realize accurate and efficient supervision.
The types of the facial shapes include round face, square face, long face, inverted triangle face, rhombus face, upper eye, lower eye, round eye, etc.
The object to be worn and identified can be a safety helmet (suitable for large-scale construction places), a mask (suitable for medical institutions, occasions where most people gather in epidemic situations, stations and the like, and also suitable for identification of the person wearing the mask), a safety belt, a helmet and the like.
Specifically, in the evaluation sorting module 303, a cosine similarity algorithm is adopted when evaluating the similarity between two faces, and the algorithm is defined as:
wherein:
in the formula, data x and y each have n multivariate attributes.
Specifically, from a geometric perspective, the cosine similarity is an included angle between two vectors; and after the included angle of the cosine similarity is obtained, converting the degree of the included angle into percentage.
In this embodiment, the application scene module 401, the tag display module 402, the feedback reminding module 403, and the supplementary update module 404 are sequentially connected through network communication; the application scene module 401 is configured to preset application scenes of the system for selection, set different operation parameters for each application scene, and train face data with different emphasis characteristics; the label display module 402 is used for framing the identity of the identified person in the monitoring video and carrying out label display on the crowd label to which the person belongs; the feedback reminding module 403 is used for feeding back to the staff in various ways when the identity of a specific crowd is recognized; the supplementary updating module 404 is used for automatically or manually supplementing and updating the face data and the related information of the new person recognized by the system for the first time into the database of the system.
The types of the application scenes comprise intelligent security, enterprise VIP customer management, unlocking access control, financial and government affair service, identity authentication, public place health and safety management, large-scale engineering site management, intelligent city service, intelligent traffic management and the like.
The crowd labels comprise VIP customers (suitable for business enterprises, building sales departments of local manufacturers and the like), famous people/actors/stars/singers (suitable for order management in public places), credit blacklists (suitable for financial institutions, transportation ticket selling and the like), inertial thieves (suitable for business monitoring, public transportation, security and protection management and the like), unworn safety helmets/masks/safety belts/helmets (suitable for construction sites, epidemic prevention and control, intelligent traffic management and the like) and the like.
The feedback reminding mode comprises a popup window, a warning sound, a short message and the like.
The embodiment also provides an operation method of the remote face recognition system, which comprises the following steps:
the method comprises the steps of firstly, building a network architecture and a system platform which comprise a processor, a database, a monitoring camera and the like, carrying out installation connection and trial operation on each basic device, selecting an application scene of the system, setting relevant limiting parameters according to the application scene, importing or inputting basic data, shooting an image video in real time through the monitoring camera and transmitting the image video to a system processing layer when the system is in operation, extracting an image containing a human face from the video by the processor, preprocessing the image, extracting human face characteristics, comparing the human face characteristics with human face data stored in the database, outputting a recognition result, carrying out mark display on the recognized human face identity or a corresponding crowd label on the video, and simultaneously feeding back the recognition result to relevant managers.
As shown in fig. 1, this embodiment further provides an exemplary product architecture of a remote face recognition system, which includes a computing processor 1, multiple face recognition technology software 2 is loaded in the computing processor 1, a face database 3 is communicatively connected outside the computing processor 1, a plurality of monitoring cameras 4 are communicatively connected to the computing processor 1 for recording monitoring videos and transmitting the monitoring videos to the processor, a display large screen 5 is connected to an external signal of the computing processor 1, and the display large screen 5 may be in a single window format or a multi-window format.
As shown in fig. 9, the present embodiment also provides an operating apparatus of a remote face recognition system, which includes a processor, a memory, and a computer program stored in the memory and operating on the processor.
The processor comprises one or more processing cores, the processor is connected with the memory through the bus, the memory is used for storing program instructions, and the processor realizes the remote face recognition system when executing the program instructions in the memory.
Alternatively, the memory may be implemented by any type or combination of volatile or non-volatile memory devices, such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disks.
Furthermore, the present invention also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the remote face recognition system.
Optionally, the present invention also provides a computer program product containing instructions which, when run on a computer, cause the computer to perform the above-described aspects of the remote face recognition system.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, and the program may be stored in a computer-readable storage medium, where the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
The foregoing shows and describes the general principles, essential features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, and the preferred embodiments of the present invention are described in the above embodiments and the description, and are not intended to limit the present invention. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (10)
1. A remote face recognition system, characterized by: comprises that
An infrastructure unit (100), a data processing unit (200), a face recognition unit (300) and an application management unit (400); the infrastructure unit (100), the data processing unit (200), the face recognition unit (300) and the application management unit (400) are sequentially connected through network communication; the infrastructure unit (100) is used for managing, controlling and distributing equipment and data bases for operation of the support system; the data processing unit (200) is used for acquiring images containing a large number of human faces from the image data and processing the images; the face recognition unit (300) is used for recognizing the face in the image through a plurality of face recognition technologies to judge the identity of the face; the application management unit (400) is used for managing the functional service of the face recognition system according to different application scenes;
the infrastructure unit (100) comprises a device management module (101), a basic database module (102), a parameter presetting module (103) and a communication support module (104);
the data processing unit (200) comprises an image shooting module (201), a lossless transmission module (202), an image management module (203) and a distribution storage module (204);
the face recognition unit (300) comprises a recognition mode module (301), a comparison recognition module (302), an evaluation sorting module (303) and a result output module (304);
the application management unit (400) comprises an application scene module (401), a label display module (402), a feedback reminding module (403) and a supplementary update module (404);
when the remote face recognition system runs, a network architecture and a system platform including a processor, a database, a monitoring camera and the like are firstly built, each basic device is installed, connected and trial run, an application scene of the system is selected, relevant limiting parameters are set according to the application scene, basic data are imported or input, when the system runs, an image video is shot and recorded in real time through the monitoring camera and transmitted to a system processing layer, the processor extracts an image containing a face from the video, preprocesses the image and extracts face features, the face features are compared with face data stored in the database, recognition results are output, the recognized face identity or a corresponding crowd label is marked and displayed on the video, and meanwhile, the recognition results are fed back to relevant managers.
2. The remote face recognition system of claim 1, wherein: the equipment management module (101), the basic database module (102) and the parameter presetting module (103) are sequentially connected through network communication; the equipment management module (101) is used for performing connection control and management on various kinds of equipment which are constructed into a network architecture supporting system operation; the basic database module (102) is used for constructing a database through a data manager so as to store face images and related information of related personnel which are pre-recorded; the parameter presetting module (103) is used for presetting label information of related personnel information and related limiting parameters in the system operation process; the communication support module (104) is used for building signal connection and data transmission channels among various devices and layers of the system through various communication means.
3. The remote face recognition system of claim 1, wherein: the signal output end of the image shooting module (201) is connected with the signal input end of the lossless transmission module (202), the signal output end of the lossless transmission module (202) is connected with the signal input end of the image management module (203), and the signal output end of the image management module (203) is connected with the signal input end of the distributed storage module (204); the image shooting and recording module (201) is used for shooting an image video in the visual field range of the monitoring camera in real time; the lossless transmission module (202) is used for transmitting the shot image data to a processing layer of the system in a real-time and lossless manner; the image management module (203) is used for intercepting images containing human faces from video images and carrying out treatment management on the images so as to identify and judge the identities of the human faces; the distribution storage module (204) is used for respectively storing and managing the original image and various types of image data generated in the processing process.
4. The remote face recognition system of claim 3, wherein: the image management module (203) comprises a graph preprocessing module (2031), a PCA dimension reduction module (2032), a feature extraction module (2033), a feature classification module (2034) and a cascade detection module (2035); the signal output end of the graph preprocessing module (2031) is connected with the signal input end of the PCA dimension reduction module (2032), the signal output end of the PCA dimension reduction module (2032) is connected with the signal input end of the feature extraction module (2033), the signal output end of the feature extraction module (2033) is connected with the signal input end of the feature classification module (2034), and the signal output end of the feature classification module (2034) is connected with the signal input end of the cascade detection module (2035); the image preprocessing module (2031) is used for intercepting a clear image containing a human face from a video image and performing processing operations such as cutting, zooming, image enhancement, color binarization and the like on the image; the PCA dimension reduction module (2032) is used for transforming the original data into a group of linearly independent representations of each dimension through linear transformation to extract main characteristic components of the data so as to realize the dimension reduction of the high-dimensional data; the feature extraction module (2033) is used for collecting a face image of a user and extracting facial feature points in the image; the feature classification module (2034) is used for placing the extracted feature points in a linear SVM for classification so as to train an SVM model which can distinguish whether the image is a face image; the cascade detection module (2035) is used for face detection and correction operation of the image through the cascade detection classifier.
5. The remote face recognition system of claim 4, wherein: in the PCA dimension reduction module (2032), the PCA dimension reduction algorithm comprises the following steps:
inputting: training sample set $ D ═ x great pocket(1)}，x{(2)}，...，xΛ{ (m) } }, a low-dimensional spatial dimension d' $;
the algorithm process is as follows:
step1, centralizing all samples, namely, performing a mean value removing operation:
step2, calculating covariance matrix XX of sampleT；
Step3, Pair covariance matrix XXTDecomposing the characteristic value;
step4, and taking the eigenvector w corresponding to the largest d' eigenvalues1，w2，...，wd′；
Step5, multiplying the original matrix by the projection matrix: X.W is the data set X' after dimensionality reduction, wherein X is m × n dimensionality, and W is [ W ═ W [, W [ ]1，w2，...，wd′]Dimension n x d';
and (3) outputting: and D, reducing the dimension of the data set X'.
6. The remote face recognition system of claim 4, wherein: in the cascade detection module (2035), the expression of a simple cascade detection classifier is as follows:
in the formula, ciRepresented by the ith weak classifier, X by the feature vector, F by the classification score, each ciRoot of HuiyouAnd outputting a classification result to the X according to the classification method of the X.
7. The remote face recognition system of claim 1, wherein: the signal output end of the identification mode module (301) is connected with the signal input end of the comparison identification module (302), the signal output end of the comparison identification module (302) is connected with the signal input end of the evaluation sorting module (303), and the signal output end of the evaluation sorting module (303) is connected with the signal input end of the result output module (304); the recognition mode module (301) is used for calling different algorithms to perform different modes of face recognition operation; the comparison and recognition module (302) is used for comparing the processed face information with a face image in a user database; the evaluation sorting module (303) is used for calculating the similarity of two faces in the acquired images and the images of the database according to percentage, and sorting a plurality of pieces of face data with similarity from high to low according to the similarity; the result output module (304) is used for outputting the face identity result with the recognition similarity reaching hundreds and the corresponding related information.
8. The remote face recognition system of claim 7, wherein: the recognition mode module (301) comprises a facial feature module (3011), a face and eye type module (3012), a face recognition module (3013) and a wearing recognition module (3014); the facial feature module (3011), the facial eye type module (3012), the facial recognition module (3013) and the wearing recognition module (3014) are sequentially connected through network communication and operate in parallel; the facial feature module (3011) is used for judging the identity of the human face through facial features accurately by detecting the location of each key point such as eye, mouth, nose outline 72 of the human face in the image; the face type eye shape module (3012) is used for carrying out key point positioning analysis on the eye shape, the eye distance and the face type of the face after the face is detected so as to obtain the facial feature data of the face type eye shape, and therefore the face identification can be more accurately judged; the facial recognition module (3013) is used for performing marking analysis on a face, acquiring key feature points of the face, accurately recognizing various facial features, and combining psychological needs of users to be suitable for drainage of marketing interaction activities; the wearing identification module (3014) is used for identifying and training data based on large-scale wearing of safety articles, matching with a field camera and combining with a face identification technology, automatically identifying the wearing condition of the safety articles of field personnel, and corresponding the result to people to realize accurate and efficient supervision.
9. The remote face recognition system of claim 7, wherein: in the evaluation sequencing module (303), a cosine similarity algorithm is adopted when the similarity of two human faces is evaluated, and the algorithm is defined as:
wherein:
in the formula, data x and y each have n multivariate attributes.
10. The remote face recognition system of claim 1, wherein: the application scene module (401), the label display module (402), the feedback reminding module (403) and the supplementary update module (404) are sequentially connected through network communication; the application scene module (401) is used for presetting application scenes of a system for selection, setting different operation parameters for each application scene, and training face data with different emphasis characteristics; the label display module (402) is used for framing the identity of the identified person in the monitoring video and marking and displaying the label of the crowd to which the person belongs; the feedback reminding module (403) is used for feeding back to staff in various ways when the identity of a specific crowd is recognized; the supplementary updating module (404) is used for automatically or manually supplementing and updating the face data and the related information of the new person which is recognized by the system for the first time into the database of the system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111645008.XA CN114694008A (en) | 2021-12-30 | 2021-12-30 | Remote face recognition system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111645008.XA CN114694008A (en) | 2021-12-30 | 2021-12-30 | Remote face recognition system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114694008A true CN114694008A (en) | 2022-07-01 |
Family
ID=82137687
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111645008.XA Pending CN114694008A (en) | 2021-12-30 | 2021-12-30 | Remote face recognition system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114694008A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116580445A (en) * | 2023-07-14 | 2023-08-11 | 江西脑控科技有限公司 | Large language model face feature analysis method, system and electronic equipment |
-
2021
- 2021-12-30 CN CN202111645008.XA patent/CN114694008A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116580445A (en) * | 2023-07-14 | 2023-08-11 | 江西脑控科技有限公司 | Large language model face feature analysis method, system and electronic equipment |
| CN116580445B (en) * | 2023-07-14 | 2024-01-09 | 江西脑控科技有限公司 | Large language model face feature analysis method, system and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109271554B (en) | Intelligent video identification system and application thereof | |
| US20220092881A1 (en) | Method and apparatus for behavior analysis, electronic apparatus, storage medium, and computer program | |
| CN111488804A (en) | Labor insurance product wearing condition detection and identity identification method based on deep learning | |
| CN109819208A (en) | A kind of dense population security monitoring management method based on artificial intelligence dynamic monitoring | |
| CN108229335A (en) | It is associated with face identification method and device, electronic equipment, storage medium, program | |
| CN105469029A (en) | System and method for object re-identification | |
| CN106156688A (en) | A kind of dynamic human face recognition methods and system | |
| CN111353338B (en) | Energy efficiency improvement method based on business hall video monitoring | |
| JPWO2017122258A1 (en) | Congestion status monitoring system and congestion status monitoring method | |
| WO2016190814A1 (en) | Method and system for facial recognition | |
| JP5785667B1 (en) | Person identification system | |
| Ancheta et al. | FEDSecurity: implementation of computer vision thru face and eye detection | |
| CN109784199A (en) | Analysis method of going together and Related product | |
| KR20190093799A (en) | Real-time missing person recognition system using cctv and method thereof | |
| CN113269091A (en) | Personnel trajectory analysis method, equipment and medium for intelligent park | |
| CN113743256A (en) | Construction site safety intelligent early warning method and device | |
| Kalunga et al. | Development of fingerprint biometrics verification and vetting management system | |
| CN115861940A (en) | Working scene behavior evaluation method and system based on human body tracking and recognition technology | |
| CN114694008A (en) | Remote face recognition system | |
| CN114581990A (en) | Intelligent running test method and device | |
| Sharma et al. | Suspect face retrieval via multicriteria decision process | |
| Pinthong et al. | The License Plate Recognition system for tracking stolen vehicles | |
| Peng et al. | [Retracted] Helmet Wearing Recognition of Construction Workers Using Convolutional Neural Network | |
| CN115908493A (en) | Community personnel track management and display method and system | |
| CN114898287A (en) | Method and device for dinner plate detection early warning, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination |