Abstract
Steganography is the primary method by which individuals can communicate covertly; cryptography, on the other hand, fails at this, as it is possible to detect (the presence of) encrypted-communication. Steganalysis has been used to detect the presence of steganography and acts as a countermeasure to it. The ongoing race between image-steganography and steganalysis methods has resulted in the need for this paper which surveys and compares developments in these two intertwined-fields. This work covers over 150 papers that demonstrate the significant improvements made in steganography and steganalysis over the last three-decades. We mention the novelty of the method proposed in each paper, as well as the evaluation results and the paper's contribution to the field. We provide several taxonomies for steganography and steganalysis methods, based on the approach and techniques underlying the methods, which allows us to perform the first comprehensive comparison of steganography and steganalysis methods. This comparison sheds light on the existing-gaps between the two connected domains and can be used to identify and prioritize the steganography methods that require immediate remediation using steganalysis methods. Lastly, we follow the chronological-evolution of steganography and steganalysis methods over the years, an overview which highlights the infinite-nature of this race.
Original language | English |
---|---|
Article number | 108711 |
Journal | Signal Processing |
Volume | 201 |
DOIs | |
State | Published - 1 Dec 2022 |
Keywords
- Deep learning
- Images
- Machine learning
- Steganalysis
- Steganography
ASJC Scopus subject areas
- Control and Systems Engineering
- Software
- Signal Processing
- Computer Vision and Pattern Recognition
- Electrical and Electronic Engineering