PVDMFWO: a novel steganographic method using PVDMF and optimization
L Akhila, V. J. Manoj · International Journal of Computers and Applications · 2025
This paper introduces an optimized steganography technique, Pixel Value Differencing using Modulus Function with Optimization (PVDMFWO), enhancing the PVDMF method. Traditional PVDMF-based methods often lack proper optimization for their constraints. Our approach models PVDMF-based steganography as a constrained optimization problem, aiming to minimize the Mean Square Error (MSE) between cover and stego images. We propose two novel approaches: Penalty Function-Based Improved Arithmetic Optimization Algorithm (PFBIAOA) and Constraint Contemplation Improved Arithmetic Optimization Algorithm (CCIAOA). PFBIAOA employs penalty functions, adding weight terms to the objective function to prevent constraint violations. In CCIAOA, a new Constraint Contemplation Algorithm (CCA) is introduced alongside IAOA to handle PVDMF method constraints. PFBIAOA achieves an average Peak Signal-to-Noise Ratio (PSNR) of 33.62 dB, Structural Similarity Index (SSIM) of 0.99, MSE of 33.58, ΔH of 0.3641, and Correlation Coefficient (CC) of 0.9884. CCIAOA outperforms existing techniques, attaining a PSNR of 39.92 dB, SSIM of 1.00, MSE of 7.51, ΔH of 0.3374, and CC of 0.9973, indicating superior image quality and imperceptibility. Both methods achieve payload capacity of 2.33 bits per pixel (bpp) and are resistant to steganalysis methods like regular-singular (RS) analysis and pixel value difference histogram (PDH) analysis. Among the two, CCIAOA yields the best performance.