Prediction-Based Lossless Data Hiding by Adaptive Coefficient-Bias Algorithm

In this paper, we propose a new loss less data hiding method by the adaptive coefficient-bias (ACB) algorithm. The proposed ACB algorithm involves two main steps: block-mean removal and pixel isolating. The block-mean removal first generates the difference blocks, which are obtained by subtracting each pixel of the host block from the predicted mean. Then, the pixel isolating minimizes distortion. However, to further dig out an extra storage, the pixel squeezing can be proceeded before pixel isolating step. Experimental results show that the quality of the marked images is good, while the payload provided by the proposed method is larger than that provided by existing techniques.