PairStruct: A Steganography Algorithm Based on Image Structure and Pair-Pixels Difference Expansion

The rapid evolution of digital technologies has heightened the need for robust data security and privacy measures. Steganography has emerged as a crucial technique for covertly embedding sensitive data into digital multimedia, such as images, while preserving visual integrity. This paper presents an improved pixel difference method for digital image steganography, leveraging the texture features of cover images to enhance imperceptibility and maintain high data recovery rates. The proposed method effectively balances payload capacity, visual quality, and data integrity by incorporating standard deviation to embed data adaptively into noisier regions. Experimental results on standard datasets demonstrate that the proposed method performs better than prior approaches. Peak Signal-to-Noise Ratio (PSNR) values consistently exceed 58 dB for payloads up to 20 kb, ensuring minimal impact on visual quality. Even at higher payloads of 100 kb, the method maintains higher PSNR values and achieves a perfect Structural Similarity Index Measure (SSIM) of 1.000 across all payload sizes. Furthermore, the proposed method achieves a 100% recovery rate, significantly outperforming prior research, which exhibits declining recovery rates at larger payloads.