Fibonacci Based Key Frame Selection and Scrambling for Video Watermarking in DWT–SVD Domain

Abstract

Video copyright is the technique for hiding sensitive data in digital video. This paper aims to introduce a new technique for video copyright namely Fibonacci based keyframe selection and scrambling for video watermarking in DWT---SVD. In this methodology, scene change detection technique is employed for identifying frequent changes in the scenes. In a specific frame, each scene is selected by using keyframes, which are generated by Fibonacci sequence. The initial seeds of the Fibonacci sequence are used as authentication key for generating keyframes. Based on the limitations of the scene changes the authentication keys are generated. The watermark is embedded only in specific frame of the changed scene in LH sub band. The secret image is scrambled using Fibonacci---Lucas transform. In embedding process, singular values (SVs) of the scrambled watermark block are added to the SVs of specific frames, produce the watermarked video collectively. In extraction phase, SVD is performed on the watermarked video to extract the scrambled watermark block from key frames. After that the extracted watermark blocks are collected together to yield the complete scrambled watermark. Later it is descrambled by using the secret keys to generate the original watermark. This method is blind since the host video is not required to extract the watermark. The experimental results show that the proposed methodology resists different image and video processing attacks. In addition, the proposed methodology improves the robustness and quality of both host video and watermark.