An Improved Histogram Based Boosting Detection Rate Video Watermarking Algorithm

Abstract

The existing histogram based video watermarking algorithm with temporal modulated is robust to combined attacks, but the watermark detection rate is not high due to watermark cannot embedded to the smoothness and still areas effectively. To increase the watermark detection rate, in this paper, we proposed the improved algorithm of shot segmentation first and then propose an improved video watermarking algorithm which firstly construct the watermark template in each frame video in the same shot through computing block based histogram and selecting the position of the relative high variance. Then we embed the watermark template into the video frame by temporal modulation without changing the destination of the shot group of the consecutive frames. The watermark sequence is extracted by comparing the correlation distribution of video frame and corresponding watermark template in the time domain. Experimental results demonstrate that the proposed algorithm is robust to recording attacks and guarantee the watermarking video quality at the same time, besides the watermark sequences can embedded to the smoothness and still areas effectively, and the watermark detection rate can increase by about 10% than previous methods.