Abstract:

In order to improve the robustness of a watermark system , theoretical analyses and experiments are carriedout to investigate the limitation and robustness in blending number of the existing iterative blending algorithm ,demonstrating that the main reason for weak robustness is the multiplying of the image error many times. Furthermore, an iterative blending algorithm based on the error correction of sub image is proposed. In this algorithm , thesub images of the watermarked host image are modified by the sub images of the original host image , and several watermark images are embedded by the control of secret keys. The proposed algorithm is of strong robustness in resistingattacks with strong relativity. Especially , to cropping attacks , the error of the extracted watermark is appro-ximately zero , thus concluding that the proposed iterative blending algorithm is more robust than the original one.

Key words: image hiding, error correction, watermark, sub image, iterative blending algorithm