Secure Visual Image Encryption Using Lorenz Chaos, Steganography, and Wavelet-Based Steganography Authentication

Medical images play a vital role in diagnosis and clinical decision-making, yet their transmission and storage pose significant privacy and security challenges. This research proposes a visual stego-encryption system that integrates the Lorenz chaotic algorithm with Discrete Wavelet Transform (DWT) to embed both secret medical data and a doctor's digital signature into a visually meaningful encrypted image (VMEI). The system employs dual-layer embedding and role-based access control, allowing administrators to input patient and doctor data while enabling doctors to perform secure validation and decryption. A series of evaluation scenarios were conducted, including variations in image resolution, geometric transformations (rotation), Lorenz initial conditions, alpha embedding parameters, and multivariate optimization, along with user-role-based validation testing. Performance metrics based on Peak Signal-to-Noise Ratio (PSNR) and Bit Error Rate (BER) demonstrate that the system consistently achieves high visual fidelity (PSNR > 30 dB) and low data loss (BER ≈ 0) across all image types. The optimal configuration—using a 4096×4096 carrier, 1024×1024 secret, and 256×256 signature with α₁ = 0.28, α₂ = 0.07, and initial condition (0.2, 0.8, 1.5)—resulted in a PSNR of 33.01 dB for the secret image. These results confirm that the proposed system provides a robust, secure, and visually accurate method for medical image encryption, suitable for integration into real-world digital healthcare infrastructures.