An Overview of Enhancing Encryption Standards for Multimedia in Explainable Artificial Intelligence Using Residue Number Systems for Security

Abstract

In spite of the increasing use of machine learning models in the applications of cyber-security, such as intrusion detection system (IDS), the vast majority of developed models are still considered black boxes. In order to improve conviction organization by permitting human specialists to grasp the basic data indication and fundamental perceptive, the use of eXplainable artificial intelligence (XAI) to interpret machine learning models has become increasingly important. Because of advancements in Internet technology and the development of efficient compression techniques, the security of digital video storage and transmission has recently gotten a lot of attention. The improvement has permitted the widespread use of video in a variety of strategies, as well as the communication of complex information, such as medical, military, and political secrets. These multimedia data are constantly subject to interception by hostile and unauthorized people all over the world when transmitted over an open network (Internet). Encryption, whether entire or selective encryption, is a frequently used and appropriate solution for tackling these security challenges. The entire video encryption strategy (also known as the Nave Approach) has been proven and demonstrated to provide a higher level of video security. However, because of its slowness in processing huge amounts of video data, it is computationally expensive and so has limited applicability in video encryption. This research proposes methods for improving multimedia encryption standards in explainable artificial intelligence using residue number systems for security.