A private watermarking for halftone image by stochastic error diffusion

In this paper we introduce a novel method to hide binary visual pattern into halftone images by stochastic error diffusion with the Jarvis kernel. And when we detect the watermarking, we only need to overlap the original halftone image without watermark and the halftone image with watermark or to subtract the two images. Then we will find the binary visual pattern directly on the overlapped or subtracted image. Simulation results show that the halftone image with watermark has good visual effect compared with the halftone image without watermark and when the halftone image with the watermark is severely stained or scratched, the detection algorithm can detect the watermark efficiently.