Secure color image encryption algorithm for face recognition using Zaslavsky and Arnold cat maps with binary bit-plane decomposition

Abstract

During airport security checks, prompt identification and data processing of face images are critical. To address this, a real-time and high-security face image encryption scheme is proposed, which leverages the Haar Cascade Classifier to precisely locate facial regions and encrypts only the identified areas.The encryption algorithm takes the identified facial region and encryption key as inputs. Using the key as a seed, the two-dimensional Zaslavsky chaotic map generates pseudo-random chaotic sequences. A sequence of encryption operations is then performed, including random noise addition, binary bit-plane decomposition, confusion, diffusion, and RGB channel obfuscation. Eventually, the processed image is output as the encrypted face image. For algorithm evaluation, image data are randomly sampled from the IMDB-WIKI dataset. Additionally, a benchmark public dataset is employed to compare the proposed scheme with state-of-the-art encryption methods. Experimental results demonstrate that the scheme exhibits superior encryption security and efficiency.