Sample Reduction-Based Pairwise Linear Regression Classification for IoT Monitoring Systems

• Published in: Applied sciences Vol. 13; no. 7; p. 4209
• Main Authors: Gao, Xizhan, Hu, Wei, Chu, Yu, Niu, Sijie
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.04.2023
• Subjects: Access control; Algorithms; anchor point; attention mechanism; Cameras; Classification; Collaboration; Data entry; Discriminant analysis; Facial recognition technology; Information storage and retrieval; Internet of Things; IoT monitoring system; large-size video; Methods; Middleware; Monitoring systems; recognition performance optimization; Regression analysis; Safety and security measures; video face recognition
• ISSN: 2076-3417; 2076-3417
• DOI: 10.3390/app13074209

At present, the development of the Internet of Things (IoT) has become a significant symbol of the information age. As an important research branch of it, IoT-based video monitoring systems have achieved rapid developments in recent years. However, the mode of front-end data collection, back-end data storage and analysis adopted by traditional monitoring systems cannot meet the requirements of real-time security. The currently widely used edge computing-based monitoring system can effectively solve the above problems, but it has high requirements for the intelligent algorithms that will be deployed at the edge end (front-end). To meet the requirements, that is, to obtain a lightweight, fast and accurate video face-recognition method, this paper proposes a novel, set-based, video face-recognition framework, called sample reduction-based pairwise linear regression classification (SRbPLRC), which contains divide SRbPLRC (DSRbPLRC), anchor point SRbPLRC (APSRbPLRC), and attention anchor point SRbPLRC (AAPSRbPLRC) methods. Extensive experiments on some popular video face-recognition databases demonstrate that the performance of proposed algorithms is better than that of several state-of-the-art classifiers. Therefore, our proposed methods can effectively meet the real-time and security requirements of IoT monitoring systems.