• Proposing a steganographic framework that hierarchically exploits cover image structure. • Provides flexibility in positioning secret data while mitigating statistical anomalies through adaptive channel selection mechanisms. • Introducing a novel concept that releases the symbolic-numeric constraint on binary digits. • Enabling the search for optimal alternative bit representations, reducing the number of bit modifications in images compared to conventional substitution methods. • Expanding the utilization of pseudo-randomness beyond mere position randomization to serve as a deviation optimization instrument. • This aims to ensure that each data insertion produces minimal visual traces and exhibits resistance to contemporary steganalysis techniques. The need for data protection is inevitable with the ease of sharing data through public networks. Existing advanced steganography methods are constrained by trade-offs caused by hand-crafted or learned distortion models. A novel initiative is presented to optimize the utilization of the cover image’s structural manipulation strategy to enhance imperceptibility. This work proposes a method that perceives binary as an alternative way to determine the data hiding decision, called ”Dual-Options”, that combines with Least Significant Bit, possibilities from random permutation, and deviation. Different sizes of cover images and random secret data are used for the experiment. Dual-Options demonstrates the ability to obtain ≥ 50% reduction of cover image binary digit changes through the embedding process. That is correlated with the value of Peak Signal-to-Noise Ratio of 80.488 dB and the Structural Similarity Index Measure of 1.0000 for the Baboon cover image on a 1-kilobit payload. This is due to the lower number of changed bits to the total image pixels, and the deviation-based equation that is used to determine the most imperceptible channel to change. The results of this work outperform existing methods without message preprocessing. Therefore, the proposed method shows the opportunity to broaden the applicability and exploration.
Zaki et al. (Sun,) studied this question.