CN113902968A - Face recognition system based on edge calculation framework - Google Patents
Face recognition system based on edge calculation framework Download PDFInfo
- Publication number
- CN113902968A CN113902968A CN202111494845.7A CN202111494845A CN113902968A CN 113902968 A CN113902968 A CN 113902968A CN 202111494845 A CN202111494845 A CN 202111494845A CN 113902968 A CN113902968 A CN 113902968A
- Authority
- CN
- China
- Prior art keywords
- face recognition
- server
- face
- registered
- recognition system
- 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
Images
Landscapes
- Collating Specific Patterns (AREA)
- Image Analysis (AREA)
Abstract
The present disclosure provides a face recognition system based on an edge computing framework, the system comprising: the system comprises a server and a face recognition model library, wherein the server is configured to store a face database and the face recognition model library, the face database comprises registered identity information and registered face feature information associated with the registered identity information, and the face recognition model library comprises at least one AI model; and the client is configured to acquire video images in real time and perform face recognition on the video images frame by frame according to the registered identity information, the registered face feature information and the AI model. By adopting the face recognition system disclosed by the invention, the high-efficiency and high-performance edge calculation can be carried out through the client, the large-scale rapid face recognition is realized, the processing pressure of a server is greatly reduced, and the flexibility and robustness of the system for large-scale application and light-weight deployment can be enhanced.
Description
Technical Field
The present disclosure relates generally to the field of data processing technologies, and in particular, to a face recognition system based on an edge computing framework.
Background
In recent years, with the continuous development of deep learning and calculation operators, the face recognition is also rapidly improved. Compared with traditional identity recognition modes such as password recognition and fingerprint recognition, the face recognition has the advantages of high processing speed, high accuracy, no contact, convenience in integration and expansion and the like, so that the face recognition method is widely used, and new application service modes such as face brushing settlement, face brushing riding, face brushing shopping and the like are promoted.
At present, a face recognition system facing various application scenes gradually shows a large-scale, lightweight and modularized development trend. For example, in the process of developing epidemic situation prevention and control towards normality and precision, large-scale rapid face recognition becomes a basic support technology, and face recognition is more and more basic and basic in cooperation with emerging applications such as body temperature detection, face recognition and trajectory analysis.
However, in the related art, the face recognition system adopts general-purpose equipment to perform full-stack development and deployment, and uses a conventional pc (personal computer) or notebook computer as an algorithm processing platform, but this method has low reliability and low cost performance, and meanwhile, the combination degree of software and hardware is limited, which is difficult to satisfy the function integration and expansion of the face recognition system.
Disclosure of Invention
In view of the above-mentioned defects or shortcomings in the related art, it is desirable to provide a face recognition system based on an edge computing framework, which can perform edge computing with high efficiency and high performance, implement large-scale fast face recognition, and enhance the flexibility and robustness of the system in response to large-scale application and light-weight deployment.
The present disclosure provides a face recognition system based on an edge computing framework, the system comprising:
the system comprises a server and a face recognition model library, wherein the server is configured to store a face database and the face recognition model library, the face database comprises registered identity information and registered face feature information associated with the registered identity information, and the face recognition model library comprises at least one AI model;
and the client is configured to acquire video images in real time and perform face recognition on the video images frame by frame according to the registered identity information, the registered face feature information and the AI model.
Optionally, in some embodiments of the present disclosure, a parent thread in the client is used to process real-time acquisition of the video image, a child thread is used to process face recognition of the video image, and data exchange is performed between the parent thread and the child thread in turn.
Optionally, in some embodiments of the present disclosure, the video images are collected in parallel between the clients, and the video images are subjected to face recognition.
Optionally, in some embodiments of the present disclosure, the server includes a main server and at least two sub servers, the main server is configured to share and synchronize data with the sub servers, wherein the sub servers are communicatively connected with at least one of the clients.
Optionally, in some embodiments of the present disclosure, the server periodically sends synchronization data and a heartbeat signal to each of the clients.
Optionally, in some embodiments of the present disclosure, the client is further configured to detect a face in the video image, and perform face registration and feature information extraction on the face;
and when the similarity value of the extracted feature information and the registered face feature information is maximum, determining the registered identity information corresponding to the registered face feature information as an identification target.
Optionally, in some embodiments of the present disclosure, the client is further configured to generate abnormal information and an early warning prompt if the identified target is an interested target, and send the abnormal information to the server for archiving.
Optionally, in some embodiments of the present disclosure, the alert prompt includes at least one of a highlight and a voice prompt.
Optionally, in some embodiments of the present disclosure, the client is further configured to continue processing the next frame of the video image if the stop identification instruction sent by the server is not received; otherwise, the connection with the server is disconnected, and the operation is finished.
Optionally, in some embodiments of the present disclosure, the server performs serial processing according to the receiving time of the exception information.
According to the technical scheme, the embodiment of the disclosure has the following advantages:
the embodiment of the disclosure provides a face recognition system based on an edge computing frame, which is characterized in that a client is linked with a server in time to perform data association mining and analysis, and face recognition is executed locally, so that the lightweight of client application and the modularization seamless connection and organic combination of the server can be realized, and the input-output intensive processing load and the computation intensive processing load in the face recognition are ingeniously distributed to the client, so that edge computing can be performed efficiently and high-performance, large-scale rapid face recognition is realized, the processing pressure of the server is greatly reduced, and the flexibility and robustness of the system for large-scale application and lightweight deployment can be enhanced.
Drawings
Other features, objects and advantages of the disclosure will become more apparent upon reading of the following detailed description of non-limiting embodiments thereof, made with reference to the accompanying drawings in which:
fig. 1 is a schematic diagram of a basic structure of a face recognition system based on an edge computing framework according to an embodiment of the present disclosure;
fig. 2 is a schematic data processing flow diagram of a face recognition system based on an edge computing framework according to an embodiment of the present disclosure;
fig. 3 is a schematic flow chart of face recognition according to an embodiment of the present disclosure;
fig. 4 is a schematic structural diagram of another face recognition system based on an edge calculation framework according to an embodiment of the present disclosure.
Detailed Description
In order to make the technical solutions of the present disclosure better understood by those skilled in the art, the technical solutions of the embodiments of the present disclosure will be clearly and completely described below with reference to the drawings in the embodiments of the present disclosure, and it is obvious that the described embodiments are only a part of the embodiments of the present disclosure, not all of the embodiments. All other embodiments, which can be derived by one of ordinary skill in the art from the embodiments disclosed herein without making any creative effort, shall fall within the scope of protection of the present disclosure.
The terms "first," "second," "third," "fourth," and the like in the description and in the claims of the present disclosure and in the drawings described above, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the disclosure described are capable of operation in sequences other than those illustrated or otherwise described herein.
Moreover, the terms "comprises," "comprising," and any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or modules is not necessarily limited to those steps or modules explicitly listed, but may include other steps or modules not expressly listed or inherent to such process, method, article, or apparatus.
For ease of understanding and explanation, the edge computing framework-based face recognition system provided by the embodiments of the present disclosure is set forth in detail below with reference to fig. 1 to 4.
Please refer to fig. 1, which is a schematic diagram illustrating a basic structure of a face recognition system based on an edge computing framework according to an embodiment of the present disclosure. The face recognition system 100 includes a server 101 and at least one client 102 communicatively connected to the server 101, wherein the server 101 is deployed in a cloud and configured to store a face database and a face recognition model library, the face database may include, but is not limited to, registered identity information and registered face feature information associated with the registered identity information, and the face recognition model library may include, but is not limited to, at least one ai (intelligent intelligence) model; the client 102 is deployed at the edge, and is configured to collect video images in real time, and perform face recognition on the video images frame by frame according to the registered identity information, the registered face feature information, and the AI model. The advantage of this arrangement is that the embodiment of the present disclosure allocates the "input/output intensive" and "computation intensive" processing loads in face recognition to the client 102, so that it can perform edge computation with high efficiency and high performance, thereby greatly reducing the processing pressure of the server 101, and implementing large-scale fast face recognition, and simultaneously enhancing the flexibility and robustness of the system in response to large-scale application and light-weight deployment.
Optionally, in some embodiments of the present disclosure, a parent thread of the client 102 is used to process real-time acquisition of a video image, a child thread is used to process face recognition of the video image, and the parent thread and the child thread perform data exchange in turn in a "ping-pong operation" manner through a custom message body, that is, the embodiments of the present disclosure implement real-time acquisition of a video image of the client and rapid face recognition by using a multithreading technique, so that processing resources can be fully utilized, and efficiency is greatly improved.
Optionally, in some embodiments of the present disclosure, video images are collected among the clients 102 in parallel, and the video images are subjected to face recognition, so that the clients 102 are independent and do not interfere with each other, and the processing efficiency is further improved.
The following describes the operation of the face recognition system 100 in detail by taking the structure shown in fig. 1 as an example and combining with the data processing flow diagram shown in fig. 2.
Specifically, when the face recognition system 100 is started, the server 101 first performs initialization, such as starting the face database and the face recognition model library in sequence, and establishes communication connection with each client 102 in sequence, and forwards initialization parameters to the client 102. Then, the server 101 not only needs to send synchronization data and heartbeat signals to each client 102 regularly, for example, the synchronization data may include, but is not limited to, data in a face database and data in a face recognition model library to implement information update and wakeup, but also needs to receive and process abnormal information sent by each client 102 at any time, so as to ensure that the entire system is in a normal operating state.
Each client 102 firstly starts a camera integrated on a local RISC-V board card, and then carries out face recognition on the video images acquired in real time frame by depending on the board card resources. Exemplarily, as shown in fig. 3, it is a schematic flow chart of face recognition provided by the embodiment of the present disclosure. At the beginning of the system operation, the client 102 needs to read all the registered facial feature information from the server 101 for use. Then, the client 102 continuously acquires video images through a camera, detects human faces in the video images, and performs face registration and feature information extraction on the human faces; and when the similarity value of the extracted feature information and the registered face feature information is maximum, determining the registered identity information corresponding to the registered face feature information as the identification target.
Further, if the target is identified as the target of interest, the client 102 further generates abnormal information and an early warning prompt, and automatically sends the abnormal information to the server 101 for archiving, for example, the early warning prompt may include but is not limited to at least one of highlighting and voice prompt. Subsequently, if the client 102 does not receive the identification stop instruction issued by the server 101, the client continues to process the next frame of video image, otherwise, the client disconnects the connection with the server 101, and ends the operation.
It should be noted that, in the embodiment of the present application, the face recognition operation of the client 102 is executed in parallel and in a loop. Only when an interested target is found, the client 102 needs to feed back the abnormal information to the server 101, and at this time, the server 101 can perform serial processing and sequential execution according to the receiving time of the abnormal information.
In addition, in an application scenario where abnormal information is not dense, the configuration of the face recognition system 100 shown in fig. 1, that is, "single server-multiple clients" can meet the requirement of the system on the overall performance. If the abnormal information is more, the processing delay of the server 101 is increased, at this time, the number of servers may be increased, and a two-stage server structure is constructed, for example, a "total server-multiple sub-servers-multiple clients" architecture formed as shown in fig. 4 is constructed, that is, the server 101 includes a total server 1011 and at least two sub-servers 1012, the total server 1011 is used for sharing and synchronizing data to each sub-server 1012, and each sub-server 1012 is in communication connection with at least one client 102, so that the overall performance of the system in large-scale application is strongly guaranteed. For each client 102, the FPGA board card based on RISC-V is adopted, so that core computing resources such as memory capacity, CPU model and the like can be configured according to actual conditions, and the cost performance and the operation reliability of the system are further integrally improved.
The embodiment of the disclosure provides a face recognition system based on an edge computing framework, wherein a client is linked with a server in time to perform data association mining and analysis, and face recognition is executed locally, so that the lightweight of client application and the modularization of the server are seamlessly connected and organically combined through a side cloud cooperation mechanism, and the processing loads of input and output intensive type and computing intensive type in the face recognition are ingeniously distributed to the client, so that edge computing can be performed efficiently and high-performance, large-scale rapid face recognition is realized, the processing pressure of the server is greatly reduced, and meanwhile, the flexibility and robustness of the system for large-scale application and lightweight deployment can be enhanced.
The above examples are only intended to illustrate the technical solutions of the present disclosure, not to limit them; although the present disclosure has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present disclosure.
Claims (10)
1. An edge computing framework based face recognition system, the system comprising:
the system comprises a server and a face recognition model library, wherein the server is configured to store a face database and the face recognition model library, the face database comprises registered identity information and registered face feature information associated with the registered identity information, and the face recognition model library comprises at least one AI model;
and the client is configured to acquire video images in real time and perform face recognition on the video images frame by frame according to the registered identity information, the registered face feature information and the AI model.
2. The edge computing framework-based face recognition system of claim 1, wherein a parent thread in the client is used for processing real-time acquisition of the video image, a child thread is used for processing face recognition of the video image, and data exchange is performed between the parent thread and the child thread in turn.
3. The edge-computing-framework-based face recognition system of claim 1, wherein the video images are collected and face recognition is performed on the video images in parallel between the clients.
4. The edge-computing framework-based face recognition system of claim 1, wherein the server comprises a main server and at least two sub-servers, the main server being configured to share and synchronize data to each of the sub-servers, wherein the sub-servers are communicatively coupled to at least one of the clients.
5. The edge-computing-framework-based face recognition system of claim 1, wherein the server periodically sends synchronization data and heartbeat signals to each of the clients.
6. The edge computing framework-based face recognition system according to any one of claims 1 to 5, wherein the client is further configured to detect a face in the video image, and perform face registration and feature information extraction on the face;
and when the similarity value of the extracted feature information and the registered face feature information is maximum, determining the registered identity information corresponding to the registered face feature information as an identification target.
7. The edge-computing-framework-based face recognition system of claim 6, wherein the client is further configured to generate exception information and an early warning prompt if the recognition target is an object of interest, and send the exception information to the server for archiving.
8. The edge computing framework-based face recognition system of claim 7, wherein the early warning prompt comprises at least one of a highlight and a voice prompt.
9. The edge-computing-frame-based face recognition system of claim 7, wherein the client is further configured to continue processing the next frame of the video image if a stop recognition instruction issued by the server is not received; otherwise, the connection with the server is disconnected, and the operation is finished.
10. The edge-computing-framework-based face recognition system of claim 7, wherein the server performs serial processing according to the receiving time of the abnormal information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111494845.7A CN113902968A (en) | 2021-12-09 | 2021-12-09 | Face recognition system based on edge calculation framework |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111494845.7A CN113902968A (en) | 2021-12-09 | 2021-12-09 | Face recognition system based on edge calculation framework |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113902968A true CN113902968A (en) | 2022-01-07 |
Family
ID=79025803
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111494845.7A Pending CN113902968A (en) | 2021-12-09 | 2021-12-09 | Face recognition system based on edge calculation framework |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113902968A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107491652A (en) * | 2017-08-30 | 2017-12-19 | 广州云从信息科技有限公司 | A kind of face off-note data analysing method based on recognition of face |
| CN111698470A (en) * | 2020-06-03 | 2020-09-22 | 河南省民盛安防服务有限公司 | Security video monitoring system based on cloud edge cooperative computing and implementation method thereof |
| US20210096911A1 (en) * | 2020-08-17 | 2021-04-01 | Essence Information Technology Co., Ltd | Fine granularity real-time supervision system based on edge computing |
| CN113065528A (en) * | 2021-05-08 | 2021-07-02 | 南京四维向量科技有限公司 | Embedded visual computing system for face recognition, counting and temperature measurement |
-
2021
- 2021-12-09 CN CN202111494845.7A patent/CN113902968A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107491652A (en) * | 2017-08-30 | 2017-12-19 | 广州云从信息科技有限公司 | A kind of face off-note data analysing method based on recognition of face |
| CN111698470A (en) * | 2020-06-03 | 2020-09-22 | 河南省民盛安防服务有限公司 | Security video monitoring system based on cloud edge cooperative computing and implementation method thereof |
| US20210096911A1 (en) * | 2020-08-17 | 2021-04-01 | Essence Information Technology Co., Ltd | Fine granularity real-time supervision system based on edge computing |
| CN113065528A (en) * | 2021-05-08 | 2021-07-02 | 南京四维向量科技有限公司 | Embedded visual computing system for face recognition, counting and temperature measurement |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20160217326A1 (en) | Fall detection device, fall detection method, fall detection camera and computer program | |
| Bolotnikova et al. | Real-time ensemble based face recognition system for NAO humanoids using local binary pattern | |
| Bistry et al. | A cloud computing approach to complex robot vision tasks using smart camera systems | |
| Lv et al. | Application of face recognition method under deep learning algorithm in embedded systems | |
| CN111614769A (en) | Intelligent behavior analysis engine system of deep learning technology and control method | |
| CN109002776B (en) | Face recognition method, system, computer device and computer-readable storage medium | |
| CN112489076A (en) | Multi-target tracking method and system | |
| CN111429476A (en) | Method and device for determining action track of target person | |
| CN106844727B (en) | Mass image characteristic data distributed acquisition processing and grading application system and method | |
| CN113902968A (en) | Face recognition system based on edge calculation framework | |
| Ghazi et al. | Embedded implementation of a deep learning smile detector | |
| Hasan et al. | Tiny head pose classification by bodily cues | |
| CN111401482A (en) | Feature point matching method and device, equipment and storage medium | |
| Berardini et al. | Edge artificial intelligence: A multi-camera video surveillance application | |
| CN116189319A (en) | Identification module, intelligent door lock and unlocking control method | |
| CN111563479B (en) | Concurrent person weight removing method, partner analyzing method and device and electronic equipment | |
| Montero et al. | Multi-Stage Dynamic Batching and On-Demand I-Vector Clustering for Cost-effective Video Surveillance. | |
| CN113886631A (en) | Video archive generation method and device and storage medium | |
| Sznaier et al. | The role of dynamics in extracting information sparsely encoded in high dimensional data streams | |
| CN111930993A (en) | Monitoring video image management platform based on big data | |
| Gaikwad et al. | Edge-based real-time face logging system for security applications | |
| CN111476616B (en) | Trajectory determination method and apparatus, electronic device and computer storage medium | |
| Nguyen et al. | Development of a Human Daily Action Recognition System for Smart-Building Applications | |
| CN116777987A (en) | Motion prediction method, motion prediction device, computer equipment and computer readable storage medium | |
| CN114283454A (en) | Training method of position relation mapping model and related device |
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 |