An efficient chaos-based image encryption algorithm using real-time object detection for smart city applications. (October 2022)

• Record Type: Journal Article
• Authors: Singh, Kedar Nath; Singh, Om Prakash; Baranwal, Naman; Singh, Amit Kumar
• Is Part Of: Sustainable energy technologies and assessments. Volume 53:Part B(2022)
• Publication Date: 2022-10
• Subjects: Digital images -- Encryption -- YOLO -- Object detection -- Security -- Time cost; Renewable energy sources -- Periodicals; Energy development -- Technological innovations -- Periodicals; Electric power production -- Periodicals; Energy storage -- Periodicals; 333.79
• DOI: 10.1016/j.seta.2022.102566
• Languages: English
• ISSNs: 2213-1388

Abstract

Abstract: Currently, an increasing number of digital images are being generated by and for the general public and professionals. Simultaneously, the protection of these images is crucial for various emerging applications. Encryption is a popular technique for protecting image data confidentiality and privacy. There are multiple existing studies on image encryption, however, most of them are computationally overhead and do not achieve high security. Considering that time cost and security have become the most important factors, designing an efficient encryption algorithm for digital images are necessary. In this paper, a secure chaos-based image encryption algorithm for smart city applications is proposed. First, the You Only Look Once v3 object detection algorithm is used to quickly identify the objects of an image. Second, initialisation and generation of a key is performed with confusion and diffusion operations to get an encrypted object image. This algorithm for encryption has the advantage of strong security performance combined with speed, which is not available in the general encryption system. Experimental results show that this proposed algorithm outperforms other state-of-the-art schemes. Graphical abstract: Highlights: A secure chaos-based image encryption algorithm for smart city applications is proposed. A neural network based object detection method is used to identify the significant object(s) of an image before encryption. Proposed algorithm is secure and efficient than similar techniques.