CN110378296A - The monitoring method and system of low network band width demand neural network based - Google Patents
The monitoring method and system of low network band width demand neural network based Download PDFInfo
- Publication number
- CN110378296A CN110378296A CN201910664304.0A CN201910664304A CN110378296A CN 110378296 A CN110378296 A CN 110378296A CN 201910664304 A CN201910664304 A CN 201910664304A CN 110378296 A CN110378296 A CN 110378296A
- Authority
- CN
- China
- Prior art keywords
- face
- neural network
- network
- recognition
- server end
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/94—Hardware or software architectures specially adapted for image or video understanding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/94—Hardware or software architectures specially adapted for image or video understanding
- G06V10/95—Hardware or software architectures specially adapted for image or video understanding structured as a network, e.g. client-server architectures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Biomedical Technology (AREA)
- Image Analysis (AREA)
Abstract
A kind of monitoring method and system of low network band width demand neural network based, by the neural network for being used for recognition of face for the neural network of Face datection, in server end deployment in test side deployment, the key frame comprising face information is obtained by carrying out Face datection to original image, recognition of face is carried out to key frame in server end and realizes monitoring.The present invention calculates pressure by disposing neural network progress aid in treatment and optimizing distribution, uses the database that smaller strip is wide and memory space is i.e. quickly to make fresh target addition detection target.
Description
Technical field
The present invention relates to a kind of technology of field of image processing, specifically a kind of low Netowrk tape neural network based
The monitoring method and system of wide demand.
Background technique
Intelligent monitor system is using image procossing, pattern-recognition and computer vision technique, by monitoring system
Increase intelligent video analysis module, fast and accurately localized accident scene, judges the abnormal conditions in monitored picture, and with most fast
Other movements are sounded an alarm or triggered with optimal mode, to effectively carry out early warning in advance, are handled in thing, subsequent evidence obtaining in time
It is full-automatic, round-the-clock, real time monitoring intelligence system.
Existing monitoring technology by the way that video flowing is directly uploaded to terminal management system, need very big bandwidth, performance compared with
High central processing unit and a large amount of memory spaces will lead to central processing unit pressure mistake when the place change for needing to monitor
Greatly, and it may require that very big network bandwidth to transmit these video informations.Simultaneously because bandwidth is needed to cause to be not easy to move,
Be inconvenient to dispose.
Summary of the invention
The present invention In view of the above shortcomings of the prior art, proposes a kind of low network band width demand neural network based
Monitoring method and system, pass through deployment neural network carry out aid in treatment and optimize distribution calculate pressure, it is wide using smaller strip
It is the database quickly to make fresh target that detection target be added with memory space.
The present invention is achieved by the following technical solutions:
The present invention relates to a kind of monitoring methods of low network band width demand neural network based, by disposing in test side
For Face datection neural network, server end deployment be used for recognition of face neural network, by original image into
Row Face datection obtains the key frame comprising face information, carries out recognition of face to key frame in server end and realizes monitoring.
The neural network for Face datection uses multitask concatenated convolutional neural network (MTCNN), the nerve
Network includes: P-NET, R-NET, O-NET；When using external neural network module, as Movidius nerve stick carries out accelerometer
When calculation, each part can be individually placed on to a neural stick cascade, to facilitate the deployment of network.
The neural network for recognition of face uses FaceNet deep learning model realization, the deep learning mould
Type uses the loss function training neural network of the maximum boundary nearest neighbour classification (LMNN) based on three sub (triplets), network
It is directly output as the vector space of 128 dimensions；The triplets includes two matching face's thumbnails and a non-matching face
Portion's thumbnail, loss function target are by distinguishing positive and negative class apart from boundary.
Technical effect
Compared with prior art, the present invention is done human face detection and recognition by way of neural network and significantly improves inspection
The accuracy of survey；By distributing Face datection and recognition of face to test side and server end, server is significantly reduced
Burden；The present invention only needs to transmit from test side to server end key picture, rather than whole section of video flowing, reduces storage demand
With the demand of network bandwidth, recognition of face network can also be made in such a way that page end adds two photos in some cases
Quickly train new target (on-line training), it can quickly addition newcomer enters database.
Detailed description of the invention
Fig. 1 is present system schematic diagram；
Fig. 2 a is embodiment Face datection result figure；
Fig. 2 b is embodiment face recognition result figure；
Fig. 3 is server state figure；
Fig. 4 is web front end surface chart；
Fig. 5 is web front end surface chart；
Fig. 6 is web front end surface chart；
Fig. 7, Fig. 8 and Fig. 9 are flow chart step by step in embodiment.
Specific embodiment
As shown in Fig. 1, Fig. 7, Fig. 8, Fig. 9, a kind of low network band width neural network based being related to for the present embodiment is needed
The monitoring system asked, wherein including: test side and the server end being attached thereto, in which: test side includes main control chip, camera shooting
Head and external neural network module；Server end include for providing the server framework and face recognition module of network service,
MTCNN network built in external neural network module is to acquire the face picture in original video, face built in face recognition module
Identification neural network (facenet) identifies face picture and default object.
The server framework is used but is not limited to: a kind of workerman (open source server framework) frame externally mentions
For web services such as http and websocket.
The present embodiment is realized especially by following manner:
Step 1, deployment system: having set up test side and server end, and confirmation test side connect good with server-side network
It is good, start test side neural network module, run the program of test side and server communication, starts server end neural network mould
Block.
Step 2, camera module acquire original video data, and main control chip is using these video datas as Face datection mind
Input through network passes to neural stick and is calculated.When detecting face, which is sent to service by main control chip
Device.Such as Fig. 7.
Step 3, server start recognition of face network after receiving picture, carry out recognition of face to picture, identification is completed
Afterwards, server records result in a manner of picture and text (current time and camera id and recognition result).
Step 4, Web show the recognition result of server to respective page, if recognition result and current needs monitoring pair
As being consistent, issuing alarm sound and popping up prompt.Such as Fig. 8.
Step 5, Web offer simultaneously make the mode of server end recognition of face network on-line training fresh target quickly update number
According to library, specific steps are as follows:
1. entering the system administration page, monitored object column is selected.
2. inputting target object name in monitor's object column in uploading pictures into base area, confirmation is clicked.
Positive face picture and the positive face picture upload of selection target object in pop-up choice box are uploaded 3. clicking.
Side face picture and the selection target object side face picture upload in pop-up choice box are uploaded 4. clicking.
After the completion of operation, server can run online training program, will be in the target newly uploaded addition database and again
Training network.As shown in Figure 9.
Above-mentioned specific implementation can by those skilled in the art under the premise of without departing substantially from the principle of the invention and objective with difference
Mode carry out local directed complete set to it, protection scope of the present invention is subject to claims and not by above-mentioned specific implementation institute
Limit, each implementation within its scope is by the constraint of the present invention.
Claims (4)
1. a kind of monitoring method of low network band width demand neural network based, which is characterized in that by being disposed in test side
For Face datection neural network, server end deployment be used for recognition of face neural network, by original image into
Row Face datection obtains the key frame comprising face information, carries out recognition of face to key frame in server end and realizes monitoring；
The neural network for Face datection uses multitask concatenated convolutional neural network；
The neural network for recognition of face uses FaceNet deep learning model realization.
2. according to the method described in claim 1, it is characterized in that, the FaceNet deep learning model use be based on three
The loss function training neural network of the maximum boundary nearest neighbour classification of son, network are directly output as the vector space of 128 dimensions；It should
Triplets includes two matching face's thumbnails and non-matching face's thumbnail, and loss function target is to pass through distance
Distinguish positive and negative class in boundary.
3. a kind of monitoring system of the low network band width demand for the neural network for realizing any of the above-described claim the method,
It is characterized in that, comprising: test side and the server end being attached thereto, in which: test side includes main control chip, camera and external
Neural network module；Server end includes for providing the server framework and face recognition module of network service, external nerve
MTCNN network built in network module is to acquire the face picture in original video, recognition of face nerve built in face recognition module
Network identifies face picture and default object.
4. system according to claim 1, characterized in that the server framework is using workerman open source service
Device frame.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910664304.0A CN110378296A (en) | 2019-07-23 | 2019-07-23 | The monitoring method and system of low network band width demand neural network based |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910664304.0A CN110378296A (en) | 2019-07-23 | 2019-07-23 | The monitoring method and system of low network band width demand neural network based |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110378296A true CN110378296A (en) | 2019-10-25 |
Family
ID=68254956
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910664304.0A Pending CN110378296A (en) | 2019-07-23 | 2019-07-23 | The monitoring method and system of low network band width demand neural network based |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110378296A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113822239A (en) * | 2021-11-22 | 2021-12-21 | 聊城中赛电子科技有限公司 | Security monitoring method and device based on electronic fence and electronic equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108564052A (en) * | 2018-04-24 | 2018-09-21 | 南京邮电大学 | Multi-cam dynamic human face recognition system based on MTCNN and method |
| US20190080154A1 (en) * | 2017-09-11 | 2019-03-14 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Integrated facial recognition method and system |
| CN109919023A (en) * | 2019-01-30 | 2019-06-21 | 长视科技股份有限公司 | A kind of networking alarm method based on recognition of face |
| CN109948568A (en) * | 2019-03-26 | 2019-06-28 | 东华大学 | Embedded human face identifying system based on ARM microprocessor and deep learning |
-
2019
- 2019-07-23 CN CN201910664304.0A patent/CN110378296A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190080154A1 (en) * | 2017-09-11 | 2019-03-14 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Integrated facial recognition method and system |
| CN108564052A (en) * | 2018-04-24 | 2018-09-21 | 南京邮电大学 | Multi-cam dynamic human face recognition system based on MTCNN and method |
| CN109919023A (en) * | 2019-01-30 | 2019-06-21 | 长视科技股份有限公司 | A kind of networking alarm method based on recognition of face |
| CN109948568A (en) * | 2019-03-26 | 2019-06-28 | 东华大学 | Embedded human face identifying system based on ARM microprocessor and deep learning |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113822239A (en) * | 2021-11-22 | 2021-12-21 | 聊城中赛电子科技有限公司 | Security monitoring method and device based on electronic fence and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109922310B (en) | Target object monitoring method, device and system | |
| CN110428522A (en) | A kind of intelligent safety and defence system of wisdom new city | |
| US9560323B2 (en) | Method and system for metadata extraction from master-slave cameras tracking system | |
| US9141184B2 (en) | Person detection system | |
| KR102203135B1 (en) | Method and system for detecting disaster damage information based on artificial intelligence using drone | |
| CN110659397B (en) | Behavior detection method and device, electronic equipment and storage medium | |
| CN108564052A (en) | Multi-cam dynamic human face recognition system based on MTCNN and method | |
| JP7064346B2 (en) | Image pickup device, information terminal, control method of image pickup device, and control method of information terminal | |
| KR101986418B1 (en) | An integrated system for searching plant diseases and insect pests | |
| CN112287741A (en) | Image processing-based farming operation management method and device | |
| TWI601425B (en) | A method for tracing an object by linking video sequences | |
| US20240046701A1 (en) | Image-based pose estimation and action detection method and apparatus | |
| US20230206093A1 (en) | Music recommendation method and apparatus | |
| CN113887318A (en) | Embedded power violation detection method and system based on edge calculation | |
| CN111126411B (en) | Abnormal behavior identification method and device | |
| CN110378296A (en) | The monitoring method and system of low network band width demand neural network based | |
| CN111695544B (en) | Information sending method and device based on crowd detection model and computer equipment | |
| KR102424098B1 (en) | Drone detection apparatus using deep learning and method thereof | |
| KR102597861B1 (en) | Monitoring apparatus and method for managing cultural assets | |
| US20230306711A1 (en) | Monitoring system, camera, analyzing device, and ai model generating method | |
| CN111291597B (en) | Crowd situation analysis method, device, equipment and system based on image | |
| CN109003414A (en) | A kind of electronic security system and working method | |
| KR102603396B1 (en) | Method and system for entity recognition and behavior pattern analysis based on video surveillance using artificial intelligence | |
| CN110378241A (en) | Crop growthing state monitoring method, device, computer equipment and storage medium | |
| KR102597859B1 (en) | Data integration apparatus and method for integrated management of cultural property-related data |
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 |