A 3D wavelet based spatial-temporal approach for video watermarking

This paper presents a watermarking procedure for digital video with spatial-temporal algorithm. Our watermarking procedure is based on the 3D wavelet transform and video scene segmentation. First, the video stream is parsed into video shots and some of shots are randomly selected to embed watermarks. By direct spread spectrum technique, a sequence of watermarks is generated from a watermark image and is adaptively added into the 3D wavelet coefficients of the video shot. The experimental result demonstrates the robustness of our video watermarking procedure to several video degradations and watermarking attacks.