CN109241111A - A kind of distributed face identification system and method for database based on memory - Google Patents
A kind of distributed face identification system and method for database based on memory Download PDFInfo
- Publication number
- CN109241111A CN109241111A CN201810981766.0A CN201810981766A CN109241111A CN 109241111 A CN109241111 A CN 109241111A CN 201810981766 A CN201810981766 A CN 201810981766A CN 109241111 A CN109241111 A CN 109241111A
- Authority
- CN
- China
- Prior art keywords
- face
- redis
- picture
- kafka
- characteristic
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Processing Or Creating Images (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention belongs to technical field of face recognition, disclose the distributed face identification system and method for a kind of database based on memory, its system includes several web cameras, several face detection modules, several face characteristic extraction modules, several face characteristic contrast modules, Redis, KAFKA, database and platform management module, and intermodule has network connection;The binaryzations such as picture stream, face characteristic value data by being forwarded by its method using middle key memory database Redis, the Key value being stored in Redis is attached in KAFKA message and is transmitted, the module for receiving message directly acquires available picture stream and characteristic value by Redis Key in KAFKA message from Redis;It is provided by the invention that The method avoids conversions required when directly being transmitted by KAFKA and inverse conversion to reduce the occupancy to Internet resources;And preservation of the data of binaryzation in memory database and extraction rate are higher than preservation and reading for KAFKA, improve the processing capacity of server.
Description
Technical field
The invention belongs to technical field of face recognition, more particularly, to a kind of distributed people of database based on memory
Face identifying system and method.
Background technique
Recognition of face is the research hotspot of current manual's intelligence and pattern-recognition, initial application derived from public security department about
The archive management and criminal investigation and case detection of criminal's photo, with the development of science and technology, rapidly and efficiently automatic recognition of face requirement is increasingly tight
Compel.The technology has good application in fields such as certificate verification, criminal investigation and case detection, in-let dimple, information security and video monitorings.
Face identification system often faces projects quantitative change more caused network during actual engineer application and takes the photograph
The dynamic of camera increases and decreases.It is higher that performance is just needed replacing when increased number of cameras is more than the server handling ability upper limit
Server will face system refitting, many maintainability problems such as migration of user data, and system branch in server replacement process
The web camera quantity held is influenced by server handling ability.
Use the system of distributed structure/architecture then can be with the web camera number of no maximum in support theory, it can be according to increase
The continuous dilatation of server hardware;Although however, distributed system in terms of dilatation and maintenance advantage it is obvious that still implementing
There is the place of many complexity.Such as the message that distributed system uses KAFKA to support as distributed message queue, KAFKA
All it is character string type, the transmission of picture, the transmission of characteristic value is carried out between process, it is necessary to be converted to Base64, user connects
Picture or feature Value Types are converted to after receiving from Base64, this mode is not only for program in conversion and inverse conversion
There is expense in performance.And it is also very big for the expense of network.
Summary of the invention
Aiming at the above defects or improvement requirements of the prior art, the present invention provides a kind of distributions of database based on memory
Formula face identification system and method use middle key using distributed structure/architecture, and by the data of the binaryzations such as picture stream, characteristic value
Memory database is forwarded, and is reduced network overhead, is improved the processing capacity of server.
To achieve the above object, according to one aspect of the present invention, a kind of distribution of database based on memory is provided
Face identification system, including several web cameras, several face detection modules, several face characteristic extraction modules, Ruo Ganren
Face Characteristic Contrast module, internal storage data library module (Redis), distributed post subscribe to message queue module (KAFKA), database
And platform management module;
Wherein, platform management module receive user setting information, generation deploy to ensure effective monitoring and control of illegal activities information, carry out web camera distribution,
Face registration and registration information storage;
Web camera module is for providing video source of the live video stream as recognition of face;
Face detection module, which is used to obtain video frame from live video stream, carries out Face datection, detects there is portrait
Video frame;And Face datection is carried out to face registration picture;
Face characteristic extraction module is used for the feature extraction carried out to the video frame with portrait, and to portrait
Face registration picture carries out feature extraction, forms registered face characteristic value collection;
Face characteristic contrast module is for comparing the face characteristic value of extraction and registered face characteristic value collection;
Database is for saving face registration information, registration feature value collection;Redis is for saving picture, face characteristic value;
Message of the KAFKA for intermodule is transmitted, and the message has the Key value of the picture, face characteristic value that are stored in Redis,
It is transmitted by the way that Key value to be attached in KAFKA message, receives the module of message by the Key value in KAFKA from Redis
It is middle to obtain the picture that can be used directly and face characteristic value.
Preferably, the distributed face identification system of above-mentioned database based on memory, face detection module will test
The video frame with portrait saved by Redis, people will be passed through with the message of video frame Redis Key value with portrait
Face detection module is sent in KAFKA queue, so that message is consumed in multiple face characteristic extraction module equilibriums.
Preferably, the distributed face identification system of above-mentioned database based on memory, face detection module are detecting
After video frame with portrait, the facial image in video frame is saved in Redis.
Preferably, the distributed face identification system of above-mentioned database based on memory, face characteristic extraction module pass through
Redis Key value obtains face picture data from Redis and therefrom extracts face characteristic, and the characteristic value of extraction is saved in
In Redis;It is sent in KAFKA queue with face characteristic value, the message of Redis Key by face characteristic extraction module,
So that message is consumed in multiple face characteristic contrast module equilibriums.
Preferably, the distributed face identification system of above-mentioned database based on memory, face characteristic contrast module pass through
Redis Key obtains face characteristic value from Redis, by the registered face characteristic value of face characteristic value and storage in the database
Collection comparison, sends platform management module by KAFKA for comparing result.
Preferably, the distributed face identification system of above-mentioned database based on memory, platform management module have specified
The function of face detection module and web camera matching relationship, for distributing face detection module for web camera, by referring to
Fixed face detection module obtains video frame from the live video stream that web camera provides and carries out Face datection.
Preferably, the distributed face identification system of above-mentioned database based on memory, platform management module is for passing through
The comparing result of face characteristic contrast module and information generation of deploying to ensure effective monitoring and control of illegal activities are deployed to ensure effective monitoring and control of illegal activities alarm, and/or will deploy to ensure effective monitoring and control of illegal activities alarm and candid photograph information pushes away
Give WEB user;Information is captured to preferably include face picture, capture panoramic pictures, capture camera and capture the time;It deploys to ensure effective monitoring and control of illegal activities letter
Breath includes video camera, specified period, specified data library and the Characteristic Contrast threshold value specified to candid photograph.
Purpose to realize the present invention, other side according to the invention provide a kind of people of database based on memory
Face recognition method,
The side that will be forwarded including picture stream, the binaryzation data of characteristic value using middle key memory database Redis
The Key value being stored in Redis is attached in KAFKA message and transmits by method, so that the module for receiving message passes through
The Redis Key value of KAFKA message institute band obtains the picture stream and characteristic value that can be used directly from Redis;
This method is required from picture or characteristic value to the conversion of Base64 when avoiding directly transmitting by KAFKA, and
From Base64 to picture or the inverse conversion of characteristic value, and the data of this kind of binaryzation of picture, characteristic value are in middle key memory number
It is higher than preservation and reading for KAFKA with extraction rate according to the preservation in the Redis of library, therefore this method can also be reduced to net
The occupancy of network resource improves the processing capacity of server.
Preferably, the face identification method of above-mentioned database based on memory, specifically comprises the following steps:
(1) it obtains video flowing and is parsed, video frame is converted into picture, Face datection is carried out to picture;
(2) to detecting that the video frame of face cuts out face picture from original image, by original panoramic picture and people
Face picture is saved in Redis;
(3) face characteristic with Redis Key corresponding to panoramic picture and facial image request message is extracted to send
It is extracted in request message queue to KAFKA face characteristic;
(4) characteristic extracting module extracts request message queue from KAFKA face characteristic and subscribes to message, works as face detection module
After generating face characteristic extraction request message, characteristic extracting module just will receive face characteristic and extract request message;
(5) Key corresponding to Redis in request message is extracted by face characteristic and obtains face picture from Redis;
(6) face characteristic extraction is carried out to face picture, the face characteristic value of extraction is saved in Redis;
(7) it will be sent with face characteristic value, face picture, the feature extraction results messages of panoramic pictures Redis Key
Into the comparison request message queue of KAFKA face characteristic;
(8) Key corresponding to Redis in request message is compared by face characteristic and obtains face characteristic value from Redis,
Characteristic value is compared with the accredited personnel's characteristic value obtained from database, sends Characteristic Contrast results messages to
In KAFKA face characteristic comparing result message queue;
(9) face characteristic comparing result is obtained from KAFKA face characteristic comparing result message queue.
Preferably, the face identification method of above-mentioned database based on memory further includes face registration step, specific as follows:
(a) registration information that user uploads, information, face picture including personnel are received;
(b) face picture is saved in database;And the face picture of upload is uploaded into Redis, face figure will be had
The Face datection request message of the Key of Redis corresponding to piece is sent in KAFKA Face datection request message queue;
(c) message is subscribed to from KAFKA Face datection request message queue, is asked when platform management module generates Face datection
Message is sought, face picture is obtained from Redis by Key corresponding to Redis in Face datection request message;
(d) to Face datection is carried out in face picture, the face that will test cuts out to be saved in Redis from picture
In;
(e) face characteristic with Redis Key corresponding to human face photo is extracted into request message and is sent to KAFKA face
In feature extraction request message queue;
(f) characteristic extracting module extracts request message queue from KAFKA face characteristic and subscribes to message, is mentioned by face characteristic
Key corresponding to Redis in request message is taken to obtain face picture from Redis;
(g) face characteristic extraction is carried out to face picture;The face characteristic value of extraction is saved in Redis;It will have
The feature extraction results messages of Redis Key corresponding to Redis Key and face picture corresponding to face characteristic value are sent to
KAFKA face characteristic extracts in results messages queue;
(h) by face characteristic extract Key corresponding to Redis in results messages obtained from Redis face picture and
Face characteristic value;
(i) face picture, portrait registration information are saved in database;Registering result information is returned to user.
In general, through the invention it is contemplated above technical scheme is compared with the prior art, can obtain down and show
Beneficial effect:
(1) distributed face identification system of database based on memory provided by the invention can be with using distributed structure/architecture
It is needed to carry out dynamic expansion according to project;The extended mode of use is a kind of soft extended mode, without appointing to original system
Business changes, and without restarting any module, only need to increase server according to actual demands of engineering, new mould is disposed on new demand servicing device
Block, modified module configuration file, starting module, module are automatically added in distributed system;Then pass through platform management module
New web camera analysis task is added to new module, i.e., dynamically completes the extension of web camera;It is entire to expand
Exhibition process is very smooth, and Operation and Maintenance is very convenient;Due to using distributed framework, the web camera number that system is supported
It, can be according to the increase continuous dilatation of server hardware according to theoretically no maximum.
(2) distributed face identification system and method for database based on memory provided by the invention, by the number of binaryzation
It is forwarded according to for example picture stream, characteristic value using middle key memory database Redis, the Key value being stored in Redis is attached
It is added in KAFKA message and is transmitted, the module for receiving message is obtained from Redis by Redis Key value in KAFKA message
Picture stream and characteristic value;
Since the KAFKA message supported all is character string type, the transmission that picture, characteristic value are carried out between process is both needed to first turn
It is changed to Base64, carries out inverse conversion after receiving using process, is converted to picture or feature Value Types from Base64;And it is of the invention
Method transmitted by the Key value in Redis, avoid and directly brought with inverse conversion by KAFKA transmission converting
Program on expense is improved by the processing capacity of server, reduces network overhead for the occupancy of Internet resources;And two
Preservation of the data of value in memory database and extraction rate are higher than preservation and reading to KAFKA, therefore use this hair
Bright method can further improve processing speed.
Detailed description of the invention
Fig. 1 is the schematic diagram of one embodiment of distributed face identification system provided by the invention;
Fig. 2 is the portrait register flow path schematic diagram in the embodiment of the present invention;
Fig. 3 is the flow diagram of the real-time video recognition of face in the embodiment of the present invention.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.As long as in addition, technical characteristic involved in the various embodiments of the present invention described below
Not constituting a conflict with each other can be combined with each other.
One embodiment of the distributed face identification system of database based on memory disclosed by the invention, including several nets
Network video camera, several face detection modules, several face characteristic extraction modules, several face characteristic contrast modules, internal storage data
Library module (Redis), distributed post subscribe to message queue module (KAFKA), MySql database and platform management module;Entirely
Portion's intermodule is communicated by network connection using TCP/IP;
Platform management module receives the information of user setting, generates information of deploying to ensure effective monitoring and control of illegal activities, and carries out web camera distribution, face note
Volume and registration information storage, distribute face detection module for web camera, are taken the photograph by specified face detection module from network
Live video stream is obtained in camera carries out Face datection;The face picture detected is saved by Redis, will have face picture
The message of Redis Key is sent in KAFKA queue by face detection module, is balanced out by multiple face characteristic extraction modules
Take message.
Face characteristic extraction module obtains face picture data from Redis and therefrom extracts face by Redis Key
The characteristic value of extraction is saved in Redis by feature.
It is sent in KAFKA queue with face characteristic value, the message of Redis Key by face characteristic extraction module,
Message is consumed by multiple face characteristic contrast module equilibriums.
Face characteristic contrast module obtains face characteristic value by Redis Key from Redis, by face characteristic value and number
According to the face characteristic collection comparison registered in library;Platform management module is sent by KAFKA by comparing result.
In a preferred embodiment, platform management module by user deploy to ensure effective monitoring and control of illegal activities information be confirmed whether generate alarm,
Alarm result information and candid photograph information are pushed to WEB user;Information is captured to include face picture, the information for capturing camera, grab
Clap the time.
Have referring to signal interaction of the Fig. 1 to each module and each intermodule in above-described embodiment, embodiment
Body description.
Headend equipment of the web camera as face identification system is the camera shooting in conjunction with traditional cameras and network technology
Machine, currently used for the web camera of recognition of face, one is the web camera based on picture stream, video camera carries face inspection
The function of survey;Another kind is the general web camera based on video flowing, and video camera provides live video stream for Face datection
Module carries out Face datection;Used in the examples is second of video camera that Face datection is carried out based on video flowing.
In the distributed face identification system that embodiment provides, on the one hand face detection module is used for the face inspection of video flowing
It surveys, is on the other hand used for the Face datection of face registration picture.Standard is mainly passed through for the Face datection of video flowing
RTSP network video stream, offline video file, mainstream network camera video stream, are converted to image data for video frame, use
Algorithm carries out the Face datection of image, whether there is face in detection image, the seat of feedback face coordinate in the picture and eyes
Then the image for including face information and face information are passed through KAFKA and Redis biography by the coordinate of mark, the coordinate of mouth, nose
It is defeated to arrive characteristic extracting module.It directly uses algorithm to carry out the Face datection of image the Face datection of face registration picture, examines
Whether have face in altimetric image, feedback face coordinate in the picture and the coordinate of eyes, the coordinate of mouth, nose coordinate;
The image for having face information and face information are transferred to characteristic extracting module by KAFKA and Redis.
The people that characteristic extracting module obtains the face information of video flow detection from KAFKA and Redis and face registration detects
Face information extracts the characteristic value of face;Characteristic value and face picture information are transferred to face characteristic by KAFKA and Redis
Contrast module and platform management module.
Face characteristic contrast module loads the face characteristic value of registration from database, by obtaining from KAFKA and Redis
The face characteristic value of video flow detection, the face characteristic value that will be loaded in the face characteristic value of live video stream detection and database
Collection is compared, and the corresponding identity information of face is determined by the comparison of characteristic value.
Management of the platform management module for other each modules in system, and for carrying out face registration, video camera is deployed to ensure effective monitoring and control of illegal activities;
Alarm of deploying to ensure effective monitoring and control of illegal activities is generated by the comparing result and information of deploying to ensure effective monitoring and control of illegal activities of face characteristic contrast module, by alarm result information and capture letter
Breath is pushed to WEB user;When candid photograph information preferably includes face picture, captures panoramic pictures, captures camera information and capture
Between.;Information of deploying to ensure effective monitoring and control of illegal activities includes video camera, specified period, specified data library and the Characteristic Contrast threshold specified to candid photograph
Value;Pass through the registered face feature of the portrait captured in specified video camera at the appointed time section and the storage of specified data library
Value collection compares, and the threshold value for reaching setting generates warning information;Information of deploying to ensure effective monitoring and control of illegal activities is issued to face spy by platform management module
Levy contrast module.
Redis is memory key-value pair data library, and pictorial information and characteristic value letter are used as in distributed system of the invention
Cease network transmission middle key.
KAFKA is that distributed post subscribes to message queue module, message sink and distribution for each intermodule, and realization disappears
Cease the load balancing of processing;Its message only supports text formatting, inconvenient for picture file, binary data transmission, need into
It just can be carried out transmission after row Base64 conversion.
MySQL database is for configuration information, face registration information and the information of deploying to ensure effective monitoring and control of illegal activities in preservation system.
The method that the distributed face identification system of database based on memory based on embodiment offer carries out recognition of face,
Being will include that picture stream, the binaryzation data of characteristic value are forwarded using middle key memory database Redis, will be stored in
Key value in Redis, which is attached in KAFKA message, to be transmitted, so that the module for receiving message can be by KAFKA message institute band
Redis Key value from obtaining workable picture stream and characteristic value directly from Redis.
Referring to Fig. 2, the method that the face identification system provided by embodiment carries out face registration includes the following steps:
(1) registration information that user is uploaded by platform management module, information, face picture including personnel are received;
(2) face picture of upload is saved in MySQL database by platform management module;
(3) face picture of upload is uploaded to Redis by platform management module, will be with corresponding to face picture
The Face datection request message of the Key of Redis is sent in KAFKA Face datection request message queue;
(4) face detection module subscribes to message from KAFKA Face datection request message queue, when platform management module produces
Stranger's face solicitation message, face detection module will receive Face datection request message;Face detection module passes through face
Key corresponding to Redis obtains face picture from Redis in solicitation message;
(5) face detection module carries out Face datection to face picture by preset Face datection algorithm, will test
Face be saved in Redis from being cut out in picture;
(6) face characteristic with Redis Key corresponding to human face photo is extracted request message hair by face detection module
KAFKA face characteristic is sent to extract in request message queue;
(7) characteristic extracting module extracts request message queue from KAFKA face characteristic and subscribes to message, works as face detection module
After generating face characteristic extraction request message, characteristic extracting module will receive face characteristic and extract request message;Feature extraction
Module extracts Key corresponding to Redis in request message by face characteristic and obtains face picture from Redis;
(8) characteristic extracting module carries out face characteristic extraction by face picture and preset face characteristic extraction algorithm;
The face characteristic value of extraction is saved in Redis;It will be right with Redis Key corresponding to face characteristic value and face picture institute
It answers the feature extraction results messages of Redis Key to be sent to KAFKA face characteristic to extract in results messages queue;
(9) platform management module extracts results messages queue from KAFKA face characteristic and subscribes to message, when face characteristic extracts
After module generates face characteristic extraction results messages, platform management module will receive face characteristic and extract results messages;
(10) platform management module is extracted Key corresponding to Redis in results messages by face characteristic and is obtained from Redis
Take face picture and face characteristic value;
(11) face picture, portrait registration information are saved in MySQL database by platform management module;It returns and infuses to user
Volume result information.
Referring to Fig. 3, in embodiment, the distributed face identification system based on the database based on memory that embodiment provides,
The method for carrying out recognition of face based on real-time grasp shoot video includes the following steps:
(1) portrait detection module obtains video flowing from web camera and is parsed, and video frame is converted to picture, leads to
It crosses preset face recognition algorithms and Face datection is carried out to image;
(2) portrait detection module is without any processing to the video frame for not detecting face;To the view for detecting face
Frequency frame cuts out face picture from original image, and original panoramic picture and face picture are saved in Redis;
(3) face detection module extracts the face characteristic with Redis Key corresponding to panoramic picture and facial image
Request message is sent to KAFKA face characteristic and extracts in request message queue;
(4) characteristic extracting module extracts request message queue from KAFKA face characteristic and subscribes to message, works as face detection module
After generating face characteristic extraction request message, characteristic extracting module just will receive face characteristic and extract request message;
(5) characteristic extracting module is extracted Key corresponding to Redis in request message by face characteristic and is obtained from Redis
Take face picture;
(6) characteristic extracting module carries out face characteristic extraction by face picture and face characteristic extraction algorithm;It will extract
Face characteristic value be saved in Redis;
(7) face characteristic extraction module will be mentioned with face characteristic value, face picture, the feature of panoramic pictures RedisKey
Results messages are taken to be sent in the comparison request message queue of KAFKA face characteristic;
(8) the personnel characteristics' value and relevant information of registration are obtained when face characteristic contrast module initializes from MySQL;People
Face Characteristic Contrast module subscribes to message from the comparison request message queue of KAFKA face characteristic, when face characteristic extraction module produces
After face characteristic compares request message, characteristic extracting module just will receive face characteristic comparison request message;
(9) face characteristic contrast module compares in request message Key corresponding to Redis from Redis by face characteristic
Middle acquisition face characteristic value, characteristic value and accredited personnel's characteristic value are compared;It sends Characteristic Contrast results messages to
In KAFKA face characteristic comparing result message queue;
(10) platform management module subscribes to message from KAFKA face characteristic comparing result message queue, when face characteristic pair
After producing face characteristic comparing result message than module, platform service module just will receive face characteristic comparing result message.
The system and method provided through the invention uses the data of the binaryzations such as picture stream, characteristic value in middle key
Deposit data library Redis is forwarded, and avoids conversion and inverse conversion required when directly transmitting by KAFKA, it is possible to reduce right
The occupancy of Internet resources improves the processing capacity of server.
As it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, not to
The limitation present invention, any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should all include
Within protection scope of the present invention.
Claims (10)
1. a kind of distributed face identification system of database based on memory, which is characterized in that if including several web cameras,
Dry face detection module, several face characteristic extraction modules, several face characteristic contrast modules, Redis, KAFKA, database and
Platform management module;
The platform management module receives user setting, generation deploy to ensure effective monitoring and control of illegal activities information, carry out web camera distribution, face registration and
Registration information storage;
Web camera module is for providing video source of the live video stream as recognition of face;
Face detection module, which is used to obtain video frame from live video stream, carries out Face datection, detects the video with portrait
Frame;And Face datection is carried out to face registration picture;
Face characteristic extraction module is used for the feature extraction carried out to the video frame with portrait, and to the face with portrait
It registers picture and carries out feature extraction, form registered face characteristic value collection;
Face characteristic contrast module is for comparing the face characteristic value of extraction and registered face characteristic value collection;
The database is for saving face registration information, registration feature value collection;Redis is for saving picture, face characteristic value;
Message of the KAFKA for intermodule is transmitted, and the message has the Key value of the picture, face characteristic value that are stored in Redis;
Key value is attached in KAFKA message and is transmitted, receives the module of message by the Key value in KAFKA message from Redis
It is middle to obtain the picture that can be used directly and face characteristic value.
2. distributed face identification system as described in claim 1, which is characterized in that the face detection module will test
The video frame with portrait saved by Redis, people will be passed through with the message of video frame Redis Key value with portrait
Face detection module is sent in KAFKA queue, so that message is consumed in multiple face characteristic extraction module equilibriums.
3. distributed face identification system as claimed in claim 1 or 2, which is characterized in that the face detection module is being examined
After measuring the video frame with portrait, the facial image in video frame is saved in Redis.
4. distributed face identification system as claimed in claim 3, which is characterized in that the face characteristic extraction module passes through
Redis Key value obtains face picture data from Redis and therefrom extracts face characteristic, and the characteristic value of extraction is saved in
In Redis;It is sent in KAFKA queue with face characteristic value, the message of Redis Key by face characteristic extraction module,
So that message is consumed in multiple face characteristic contrast module equilibriums.
5. distributed face identification system as described in claim 3 or 4, which is characterized in that the face characteristic contrast module
Face characteristic value is obtained from Redis by Redis Key, the registered face of face characteristic value and storage in the database is special
The comparison of value indicative collection, sends platform management module by KAFKA for comparing result.
6. distributed face identification system as claimed in claim 1 or 2, which is characterized in that the platform management module also has
There is the function of specified detection module Yu web camera matching relationship, for distributing face detection module for web camera, by
Specified face detection module obtains video frame from the live video stream that web camera provides and carries out Face datection.
7. distributed face identification system as claimed in claim 1 or 2, which is characterized in that the platform management module is used for
Alarm of deploying to ensure effective monitoring and control of illegal activities is generated by the comparing result and information of deploying to ensure effective monitoring and control of illegal activities of face characteristic contrast module, and/or will deploy to ensure effective monitoring and control of illegal activities alarm and candid photograph letter
Breath is pushed to WEB user.
8. a kind of face identification method of database based on memory, which is characterized in that will include picture stream, the binaryzation of characteristic value
The Key value being stored in Redis is attached to KAFKA by the method that data are forwarded using middle key memory database Redis
It is transmitted in message, so that the module for receiving message is obtained from Redis by the Redis Key value of KAFKA message institute band
The picture stream and characteristic value that can be used directly, to avoid conversion and inverse conversion required when directly being transmitted by KAFKA.
9. a kind of face identification method of database based on memory, which comprises the steps of:
(1) it obtains video flowing and is parsed, video frame is converted into picture, Face datection is carried out to picture;
(2) to detecting that the video frame of face cuts out face picture from original image, by original panoramic picture and face figure
Piece is saved in Redis;
(3) face characteristic with Redis Key corresponding to panoramic picture and facial image request message is extracted to be sent to
KAFKA face characteristic extracts in request message queue;
(4) characteristic extracting module extracts request message queue from KAFKA face characteristic and subscribes to message, when face detection module generates
After face characteristic extracts request message, characteristic extracting module just will receive face characteristic and extract request message;
(5) Key corresponding to Redis in request message is extracted by face characteristic and obtains face picture from Redis;
(6) face characteristic extraction is carried out to face picture, the face characteristic value of extraction is saved in Redis;
(7) it will be sent to face characteristic value, face picture, the feature extraction results messages of panoramic pictures Redis Key
KAFKA face characteristic compares in request message queue;
(8) Key corresponding to Redis in request message is compared by face characteristic and obtains face characteristic value from Redis, it will be special
Value indicative is compared with the accredited personnel's characteristic value obtained from database, sends KAFKA people for Characteristic Contrast results messages
In face Characteristic Contrast results messages queue;
(9) face characteristic comparing result is obtained from KAFKA face characteristic comparing result message queue.
10. a kind of face identification method of database based on memory, which is characterized in that further include face registration, including walk as follows
It is rapid:
(a) registration information that user uploads, information, face picture including personnel are received;
(b) face picture is saved in database;And the face picture of upload is uploaded into Redis, face picture institute will be had
The Face datection request message of the Key of corresponding Redis is sent in KAFKA Face datection request message queue;
(c) message is subscribed to from KAFKA Face datection request message queue, disappeared when platform management module generates Face datection request
Breath, obtains face picture by Key corresponding to Redis in Face datection request message from Redis;
(d) to Face datection is carried out in face picture, the face that will test is saved in Redis from cutting out in picture;
(e) face characteristic with Redis Key corresponding to human face photo is extracted into request message and is sent to KAFKA face characteristic
It extracts in request message queue;
(f) characteristic extracting module extracts request message queue from KAFKA face characteristic and subscribes to message, is asked by face characteristic extraction
Key corresponding to Redis in message is asked to obtain face picture from Redis;
(g) face characteristic extraction is carried out to face picture;The face characteristic value of extraction is saved in Redis;Face will be had
The feature extraction results messages of Redis Key corresponding to Redis Key and face picture corresponding to characteristic value are sent to KAFKA people
In face feature extraction results messages queue;
(h) Key corresponding to Redis in results messages is extracted by face characteristic and obtains face picture and face from Redis
Characteristic value;
(i) face picture, portrait registration information are saved in database;Registering result information is returned to user.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810981766.0A CN109241111B (en) | 2018-08-27 | 2018-08-27 | Distributed face recognition system and method based on memory database |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810981766.0A CN109241111B (en) | 2018-08-27 | 2018-08-27 | Distributed face recognition system and method based on memory database |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109241111A true CN109241111A (en) | 2019-01-18 |
CN109241111B CN109241111B (en) | 2020-10-23 |
Family
ID=65069251
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810981766.0A Active CN109241111B (en) | 2018-08-27 | 2018-08-27 | Distributed face recognition system and method based on memory database |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109241111B (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110111458A (en) * | 2019-04-16 | 2019-08-09 | 深圳市多度科技有限公司 | Realize method, system and door access machine that face is deployed to ensure effective monitoring and control of illegal activities |
CN110674757A (en) * | 2019-09-25 | 2020-01-10 | 北京旷视科技有限公司 | Deployment control method, system and computer readable storage medium |
CN111159136A (en) * | 2019-12-27 | 2020-05-15 | 中山大学 | Safety management system and method for face picture management and synchronization |
CN111209119A (en) * | 2020-01-21 | 2020-05-29 | 成都国翼电子技术有限公司 | Load balancing method for face snapshot rifle bolt |
CN111291628A (en) * | 2020-01-17 | 2020-06-16 | 黄芸芸 | Face data distributed recognition and storage architecture based on block chain technology |
CN111314350A (en) * | 2020-02-19 | 2020-06-19 | 深圳中兴网信科技有限公司 | Image storage system, storage method, calling system and calling method |
CN111431699A (en) * | 2019-09-29 | 2020-07-17 | 杭州海康威视数字技术股份有限公司 | Method, device and system for quickly validating face authentication function |
CN111753743A (en) * | 2020-06-28 | 2020-10-09 | 武汉虹信技术服务有限责任公司 | Face recognition method and system based on gatekeeper |
CN111950347A (en) * | 2020-06-29 | 2020-11-17 | 武汉烽火众智数字技术有限责任公司 | Control personnel early warning system and method based on face recognition |
CN112700214A (en) * | 2020-12-29 | 2021-04-23 | 创优数字科技(广东)有限公司 | Face attendance data processing method and device, server and storage medium |
CN112966642A (en) * | 2021-03-23 | 2021-06-15 | 云账户技术(天津)有限公司 | Face recognition method, system and stream processing platform |
CN113051045A (en) * | 2019-12-27 | 2021-06-29 | 南京甄视智能科技有限公司 | Method and device for dynamically balancing equipment end load under IOT platform |
CN113254686A (en) * | 2021-04-02 | 2021-08-13 | 青岛以萨数据技术有限公司 | Personnel behavior detection method, device and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20170075721A1 (en) * | 2015-09-16 | 2017-03-16 | Salesforce.Com, Inc. | Providing strong ordering in multi-stage streaming processing |
US20170161550A1 (en) * | 2015-12-03 | 2017-06-08 | Nec Corporation Of America | System and method for capturing images used in facial recognition through effective use of exposure management |
CN107104961A (en) * | 2017-04-21 | 2017-08-29 | 中国电子科技集团公司第二十八研究所 | Distributed real-time video monitoring processing system based on ZooKeeper |
CN108390881A (en) * | 2018-02-27 | 2018-08-10 | 北京焦点新干线信息技术有限公司 | A kind of distribution high concurrent real-time messages method for pushing and system |
-
2018
- 2018-08-27 CN CN201810981766.0A patent/CN109241111B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20170075721A1 (en) * | 2015-09-16 | 2017-03-16 | Salesforce.Com, Inc. | Providing strong ordering in multi-stage streaming processing |
US20170161550A1 (en) * | 2015-12-03 | 2017-06-08 | Nec Corporation Of America | System and method for capturing images used in facial recognition through effective use of exposure management |
CN107104961A (en) * | 2017-04-21 | 2017-08-29 | 中国电子科技集团公司第二十八研究所 | Distributed real-time video monitoring processing system based on ZooKeeper |
CN108390881A (en) * | 2018-02-27 | 2018-08-10 | 北京焦点新干线信息技术有限公司 | A kind of distribution high concurrent real-time messages method for pushing and system |
Cited By (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110111458A (en) * | 2019-04-16 | 2019-08-09 | 深圳市多度科技有限公司 | Realize method, system and door access machine that face is deployed to ensure effective monitoring and control of illegal activities |
CN110674757A (en) * | 2019-09-25 | 2020-01-10 | 北京旷视科技有限公司 | Deployment control method, system and computer readable storage medium |
CN110674757B (en) * | 2019-09-25 | 2022-07-19 | 北京旷视科技有限公司 | Deployment control method, system and computer readable storage medium |
CN111431699A (en) * | 2019-09-29 | 2020-07-17 | 杭州海康威视数字技术股份有限公司 | Method, device and system for quickly validating face authentication function |
CN111159136A (en) * | 2019-12-27 | 2020-05-15 | 中山大学 | Safety management system and method for face picture management and synchronization |
CN113051045A (en) * | 2019-12-27 | 2021-06-29 | 南京甄视智能科技有限公司 | Method and device for dynamically balancing equipment end load under IOT platform |
CN111159136B (en) * | 2019-12-27 | 2023-10-17 | 中山大学 | Face picture management and synchronization safety management system and method |
CN113051045B (en) * | 2019-12-27 | 2022-08-16 | 南京甄视智能科技有限公司 | Method and device for dynamically balancing equipment end load under IOT platform |
CN111291628B (en) * | 2020-01-17 | 2024-02-06 | 黄芸芸 | Face data distributed identification and storage architecture based on block chain technology |
CN111291628A (en) * | 2020-01-17 | 2020-06-16 | 黄芸芸 | Face data distributed recognition and storage architecture based on block chain technology |
CN111209119A (en) * | 2020-01-21 | 2020-05-29 | 成都国翼电子技术有限公司 | Load balancing method for face snapshot rifle bolt |
CN111314350A (en) * | 2020-02-19 | 2020-06-19 | 深圳中兴网信科技有限公司 | Image storage system, storage method, calling system and calling method |
CN111753743A (en) * | 2020-06-28 | 2020-10-09 | 武汉虹信技术服务有限责任公司 | Face recognition method and system based on gatekeeper |
CN111753743B (en) * | 2020-06-28 | 2023-05-19 | 武汉虹信技术服务有限责任公司 | Face recognition method and system based on gatekeeper |
CN111950347A (en) * | 2020-06-29 | 2020-11-17 | 武汉烽火众智数字技术有限责任公司 | Control personnel early warning system and method based on face recognition |
CN112700214A (en) * | 2020-12-29 | 2021-04-23 | 创优数字科技(广东)有限公司 | Face attendance data processing method and device, server and storage medium |
CN112966642A (en) * | 2021-03-23 | 2021-06-15 | 云账户技术(天津)有限公司 | Face recognition method, system and stream processing platform |
CN113254686A (en) * | 2021-04-02 | 2021-08-13 | 青岛以萨数据技术有限公司 | Personnel behavior detection method, device and storage medium |
CN113254686B (en) * | 2021-04-02 | 2023-08-01 | 青岛以萨数据技术有限公司 | Personnel behavior detection method, device and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN109241111B (en) | 2020-10-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109241111A (en) | A kind of distributed face identification system and method for database based on memory | |
CN111698470B (en) | Security video monitoring system based on cloud edge cooperative computing and implementation method thereof | |
CN101895727B (en) | Monitoring system, image capturing apparatus, analysis apparatus, and monitoring method | |
CN109492536B (en) | Face recognition method and system based on 5G framework | |
US8600121B2 (en) | Face recognition system and method | |
IL258817A (en) | Methods and apparatus for false positive minimization in facial recognition applications | |
US20120207356A1 (en) | Targeted content acquisition using image analysis | |
CN110458154A (en) | Face identification method, device and computer readable storage medium | |
JP2022048147A5 (en) | ||
CN110263613A (en) | Monitor video processing method and processing device | |
CN108364374A (en) | Face access control device based on deep learning and method | |
CN104581062A (en) | Video monitoring method and system capable of realizing identity information and video linkage | |
CN109711318A (en) | A kind of plurality of human faces detection and tracking based on video flowing | |
Gupta et al. | CrowdVAS-Net: A deep-CNN based framework to detect abnormal crowd-motion behavior in videos for predicting crowd disaster | |
CN109658040A (en) | Method, apparatus, equipment and the computer storage medium of conference management | |
CN104079552A (en) | Authentication system and method of operating the same | |
CN108846912A (en) | Work attendance method, terminal and server | |
CN109840962A (en) | Recognition of face movable attendance checking system and method | |
CN109961031A (en) | Face fusion identifies identification, target person information display method, early warning supervision method and system | |
CN111586432B (en) | Method and device for determining air-broadcast live broadcast room, server and storage medium | |
CN111918039A (en) | Artificial intelligence high risk operation management and control system based on 5G network | |
US20190147251A1 (en) | Information processing apparatus, monitoring system, method, and non-transitory computer-readable storage medium | |
KR102299960B1 (en) | Apparatus and method for recommending keyword related to video | |
CN111464819A (en) | Live image detection method, device, equipment and storage medium | |
CN109684961A (en) | Population management's method and device based on recognition of face |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |