Multilevel Reversible Video Watermarking with Improvement on Occluded Areas
İbrahim Yıldırım · Sakarya University Journal of Science · 2018
This paper investigates the influence of the occlusion in reversible video watermarking technique based on histogram modification of motion compensated prediction error and presents a novel reversible video watermarking algorithm which has better embedding capacity and PSNR in occluded areas. In occluded areas, conventional motion estimation techniques calculate higher prediction error. Therefore, watermarking algorithm which exploits prediction error has small embedding capacity and poor image quality. To reduce the prediction error, occlusion detection algorithms must be employed in motion estimation.