MOLSBM: A Multi-Objective LSB Matching Steganography Method
Vajiheh Sabeti, Sepideh Faiazi · 2024
A simple steganography technique called Least Significant Bit Matching (LSBM) has been vulnerable to detection in a number of different attacks. Imperceptibility (maintaining high perceptual image quality) and security are important steganography features. By using a better method, this research seeks to improve security as well as imperceptibility. Unlike most conventional steganography techniques, which concentrate on single-objective optimization, this work simultaneously optimizes security and imperceptibility using the Non-Dominated Sorting Genetic Algorithm II (NSGA-II). The cover image is divided into blocks by the suggested method, and each block has two main decisions to make: (1) selecting the seed for the pseudo-random number generator to find the best pixels for data embedding, and (2) selecting whether to adjust the pixel value if there is a discrepancy between the data bit and pixel LSB. Optimal pixels have the highest data bit-LSB match, and pixel value adjustments are made to minimize block histogram variations. NSGA-II is used for this multi-objective optimization. Comparative analysis shows that the proposed method significantly improved image quality metrics and reduced detection accuracy across various embedding rates. The Peak Signal-to-Noise Ratio (PSNR) was approximately 57.85, 55.85, and 53.15 at embedding rates of 0.3,0.5, and 0.8 bpp, respectively, reflecting a $\mathbf{2 - 3} \%$ improvement over LSBM Along with reducing the likelihood of detection by attacks.