Artificial Intelligence-Based Security Authentication: Applications in Wireless Multimedia Networks

Abstract

Security provisioning has become a key issue in wireless multimedia networks due to their vital roles in supporting various services. Conventional security solutions have deficiencies in computing efficiency and addressing emerging security challenges. In addition, traditional upper-layer authentication ignores the protection of the physical layer, leading to severe privacy data leakage. In this paper, we envision a new deep learning (DL)-enabled security authentication scheme to overcome these issues, while implementing Blind Feature Learning (BFL) and Lightweight Physical Layer Authentication (LPLA). Specifically, an intelligent authentication method is developed by exploring neural networks at the data collection unit to learn the characteristics of data. Then we propose a holistic authentication scheme based on machine learning to identify malicious multimedia devices. Experimental analysis verifies that the proposed scheme can guarantee the privacy of wireless multimedia sensors and achieve lightweight authentication. Performance results indicate that artificial intelligence-enabled authentication scheme improves the overall security of multimedia networks.

Keywords

• Multimedia network
• machine learning
• security
• neural network