A Lightweight Secure IoT Surveillance Framework Based on DCT-DFRT Algorithms

Abstract

In this paper, we propose an energy-efficient surveillance framework for real-time video processing. The proposed framework guarantees the confidentiality of important frames and transfers them for a real-time decision. First, we extract the keyframes from the surveillance video using a lightweight summarization technique based on a fast histogram-clustering approach. Then, we employ an enhanced discrete cosine transform (DCT) compression technique to reduce the size of the extracted key frames. Finally, the cryptosystem encrypts these keyframes using a lightweight image encryption scheme based on discrete fractional random transform (DFRT) and Chen chaotic system. The proposed framework is fast and ensures real-time processing. Furthermore, this framework has the ability to reduce the transmission cost, and storage required during transmitting the video surveillance.