Uncovering Hidden Data a Comparative Study of Statistical Detection Models for Grayscale Image Steganography

With the widespread use of digital images in communication and storage, detecting hidden information embedded through steganography has become increasingly important for information security. This review presents a comparative analysis of three classical statistical steganalysis techniques - Histogram-based method, RS (Regular-Singular) analysis, and Chi-square test - applied to grayscale images. Each method's underlying principles, sensitivity to embedding rates, computational complexity, and robustness are systematically discussed. Experimental results on BMP image datasets with varying embedding rates highlight their respective strengths and limitations, including detection accuracy, processing efficiency, and error characteristics. The review also explores the adaptability of these methods to different embedding scenarios and potential improvements through multi-scale analysis and hybrid approaches. This work aims to provide researchers and practitioners with a comprehensive understanding of foundational statistical steganalysis methods and to guide future developments in image steganography detection.