Frequency Analysis Attack for Mapping-Based Coverless Image Steganography
Year: 2026
Type: conference-paper
Abstract: Existing steganalysis algorithms were specifically designed to detect minor modification to incurious images so that the existing embedding-based steganography methods need to be improved to survive the detection of the passive warden. The coverless image steganography directly uses the cover images as the stego-images and is considered to be secure to steganalysis algorithms. However, due to the limitation of the embedding capacity, coverless steganography methods require a large number of stego-images to transmit a secret text, which gives passive wardens the opportunity to discover the existence of covert text communication. In this paper, based on the characteristic of unbalanced letter frequencies in English text, we propose a frequency analysis attack that can be used to detect the existence of covert text transmission using mapping-based coverless image steganography. With the threat of frequency analysis attack, more secure structure need to be included in the existing mapping-based steganography methods to avoid detection of passive wardens for applications of secure text communications.
Authors: Chiang-Lung Liu, Jo-Yen Nieh, Shao-Hung Cheng
Institution: National Defense University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7204113341