Temporal Statistic Based Video Watermarking Scheme Robust against Geometric Attacks and Frame Dropping

  • Authors:

  • Chong Chen;Jiangqun Ni;Jiwu Huang

  • Affiliations:

  • School of Information Science and Technology, Sun Yat-Sen University, Guangzhou, P.R. China 510275;School of Information Science and Technology, Sun Yat-Sen University, Guangzhou, P.R. China 510275;School of Information Science and Technology, Sun Yat-Sen University, Guangzhou, P.R. China 510275

  • Venue:
  • IWDW '09 Proceedings of the 8th International Workshop on Digital Watermarking
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Robustness against geometric attacks is still considered to be one of the major challenging issues in the development of watermarking algorithm, especially for video watermarking. This paper presents a multi-bit video watermarking scheme resilient to geometric attacks and frame dropping by utilizing the geometric invariant sequence generated from the average dc energy of the input video and the temporal statistic described by its histogram. Firstly, the average dc energy of frame is introduced and three of its geometric invariance natures, i.e. rotation, scaling and translation, are proved. Then the watermarks are embedded into the histogram of average dc energy sequence of video frame. The embedding strategy is implemented by reassigning the populations in groups of two consecutive bins in the histogram. The simulation results demonstrate that the proposed watermarking scheme achieves satisfactory performances against geometric and common signal processing attacks, and intense frame dropping and averaging.