A Comprehensive Survey of Digital Image Steganography and Steganalysis
Weiqi Luo, Kangkang Wei, Qiushi Li, Miaoxin Ye, Shunquan Tan, Weixuan Tang, Jiwu Huang
Abstract
In the realm of digital communications, steganography and steganalysis have become a solution for securely exchanging covert information. This survey initiates with an exploration of the widely used passive-warden scenario model, analyzing its significance, key performance indicators, relevant databases, and clarifying some commonly misunderstood fundamental concepts associated with this model. Subsequently, the paper comprehensively examines the evolution and current state of digital image steganography and ste-ganalysis, highlighting the transition from traditional handcrafted based methods to sophisticated deep learning based techniques developed over the past two decades. It offers thorough descriptions and evaluations of typical methods in both steganography and steganalysis, with a particular emphasis on deep learning-based techniques that have emerged in recent years. Furthermore, the survey identifies significant challenges currently faced in translating theoretical research into practical applications. By integrating these insights, the survey not only charts the historical development and technological advancements in steganography and steganalysis but also establishes a proactive agenda for future research aimed at enhancing security in covert communications.