Images Encryption and Transmission for Neuromorphic Systems Based on Chaotic Dynamic Behaviors
Yucong Xia, Siyao Zhang, Huamin Wang · 2024
In neuromorphic computing, robust image encryption is crucial for safeguarding sensitive data. This paper introduces an innovative one-dimensional logistic-sine chaotic map that generates highly unpredictable and resistant-to-statistical-analysis sequences, making them difficult to decipher. In the proposed image encryption algorithm, we employ a dynamic key generation algorithm to generate encryption keys, thereby reducing the risk of key leakage or reuse. Furthermore, by utilizing block inter-plane scrambling and dynamic pixel diffusion, our image encryption algorithm achieves a high level of encryption quality. Notably, the Discrete Wavelet Transform is used to embed the cipher image into a carrier image, reducing the risk of interception during transmission. Extensive experimentation and evaluation on specific scene images demonstrate the excellent performance of the proposed image encryption algorithm in terms of security and confidentiality: the devised encryption scheme attains an NPCR of 99.6197% and UACI of 33.2966%, along with an average entropy of 7.9994, making it highly suitable for neuromorphic applications.