Content Adaptive Watermarking Based on Regions Segmentation

This paper describes a new method of watermarking based on visual contents and using four spaces insertion of insertion to take advantage of properties of each of them. This is realized initially by segmenting image in unrefined regions and then adapting every space of insertion to these regions in order to maximize the invisibility of the marked image. Experimental results show that the watermark is robust to several distortions including geometric alterations, JPEG coding at different qualities, STIRMA4RK attack, filtering and cropping.