Enhanced adaptive data hiding method using LSB and pixel value differencing

High embedding capacity is considered as one of the significant challenges in information hiding. In the spatial domain, most of the high capacity based strategies employed the least significant bit (LSB) and pixel value differencing (PVD) techniques. However, these are suffered from the incorrect recovery of secret data, inefficient utilization of pixel intensities, and limited usage of pixel difference range table. In this paper, a new solution is proposed to alleviate the shortcoming by exploiting the pixel intensities ranges for high embedding rates while achieving the 100% recovery of secret data in the extraction phase. The proposed method divides the pixels into non-overlapping pixels block; further, compute the pixels differences along with respective ranges to determine the size of secret bits for the underlined adaptive LSB embedding process. In the extensive experiments, the proposed scheme outperforms the embedding capacity and maintains acceptable visual imperceptibility at various embedding rates, and also shows successfully the resistance against RS analysis and machine learning-based detection attacks.

Automated summary

A new solution is proposed to address the high embedding capacity challenge in information hiding by exploiting pixel intensity ranges to increase embedding rates while achieving 100% recovery of secret data. The proposed method outperforms in embedding capacity and maintains acceptable visual imperceptibility at various embedding rates, while successfully resisting RS analysis and machine learning-based detection attacks.