Video Watermarking based on Scene Detection and 3D DFT

Y. Liu (Canada), D. Wang (USA), and J. Zhao (Canada)

Keywords

Digital video watermarking 3D DFT Scene detection Histogram MPEG-4

Abstract

This paper presents a new watermarking approach to video contents protection. This scheme comprises three portions: scene detection, watermark embedding, and watermark detection. The scene detec tion is applied to divide a video sequence into different scenes before watermark embedding. The histogram difference is used for scene change detection. The randomly generated watermark sequences are embedded in the magnitude of the three-dimensional Discrete Fourier Transform (3D DFT) of each scene group. The blue channel is the only channel we consider for embedding based on the fact that human visual system is less sensitive to this component. The watermark em bedding location is chosen in the middle frequency of the DC frame in the 3D DFT magnitude. The experimental results show that this method is robust to frame loss, MPEG-4 compression, frame swap ping, frame rate changing, aspect ratio conversion, noise addition and ļ¬ltering.

Important Links:



Go Back