Reversible Data Hiding for JPEG Images Based on Point Distortion Estimation Optimization

Abstract

Reversible data hiding (RDH) for joint photographic experts group (JPEG) images has been extensively studied to achieve a perfect balance between image visual quality and file size increment. In this paper, we propose RDH algorithm for JPEG images, optimized through point distortion estimation (PDEO). First, we construct a PDEO model that takes into account block smoothness, quantization step, and zigzag sorting position, and then arrange each AC coefficient in ascending order according to the point distortion estimate, because preferentially embedding in the AC coefficient with less distortion can obtain better image quality. Second, we design the principle of "pair when ±1 appears" to makes full use of ±1 coefficients to increase the embedding capacity. Finally, the secret data is embedded into the AC coefficient pairs by using our designed 1D and 2D histogram mapping generation strategy. Experimental results demonstrate that the proposed algorithm offers superior visual image quality compared to other advanced JPEG RDH algorithms, with a smaller increase in file size and a more optimal balance in runtime complexity.