OSMH: An Optimized Steganography Model for Healthcare CT Images Using AES Encryption and Adaptive Huffman Coding

Abstract

Steganography provides a covert mechanism for embedding sensitive data within a carrier medium, such as a digital image, while maintaining its visual integrity. This paper proposes an advanced adaptive Steganography technique that combines AES-128 encryption for security, Adaptive Huffman compression for payload efficiency, and a progressive multi-bit embedding strategy to conceal text messages within RGB images. The method processes a cover image (e.g., a 512×512 RGB image) and a text message (up to 500 characters), encrypting it with AES-128 using a 16-byte key, compressing it with Adaptive Huffman coding, and embedding it into edge-detected "unimportant" pixels across red, green, and blue channels in steps from 1 to 8 bits per pixel. For each step, Mean Squared Error (MSE) and Peak Signal-to-Noise Ratio (PSNR) are calculated, printed to the console, visualized in subplots, and logged to a text file, ensuring a detailed quality assessment. Experimental results demonstrate imperceptibility (PSNR > 40 dB) across all steps, with the progressive approach outperforming traditional LSB methods in flexibility, security, and capacity. This framework offers a robust, practical solution for secure communication, balancing distortion, payload size, and computational efficiency.