Nonlinear 4D Hyper - Chotic Dynamics and Invariant Spatial Steganography for Secure Visual Information Transmission

Abstract

The rapid growth of public network communications has resulted in an unprecedented. increase in multimedia data transmission, requiring good security. Conventional single-layer cryptographic and steganographic techniques often struggle to balance efficiency, payload capacity, and imperceptibility, payload size, and imperceptibility. To solve these issues, this work presents a safe, dual-layer visual information transfer method using nonlinear physical dynamics. Initial security uses a 4D continuous hyper-chaotic confusion-diffusion engine. The two positive Lyapunov exponents make this dynamical system sensitive to initial circumstances and have a key space of about 2450, too big for exhaustive search assaults. The innovative adaptive Least Significant Bit (LSB) steganographic second layer uses simply an invariant edge map from the top 5 MSBs. This invariant map allocates image-dependent capacity of 3 bits in difficult high-frequency areas and 1 bit in smooth homogenous areas. Based on thorough testing, this method increases payload by 9.72% over standard global LSB methods without compromising visual integrity. Information entropy (∼7.999), NPCR (∼99.61%), and UACI (∼34.01%) demonstrate the framework's exceptional cryptographic performance The suggested adaptive steganography achieves PSNRs above 50 dB and SSIMs above 0.99 in all situations, averaging 53.36 dB and 0.99. Thus, the suggested adaptive steganography achieves a Peak Signal-to-Noise Ratio (PSNR) of over 50 dB and a Structural Similarity Index (SSIM) of above 0.99 in all situations, with an average of 53.36 dB and 0.99. Critical communications can be undetected, mathematically sound, and high-capacity using the two-layered approach.