A robust video watermarking technique based on key-frames extraction and contourlet transform

Abstract

Video watermarking has encountered many types of attacks that can have some affection on the embedded watermark. Security, strength, and reliability are the essential challenges for any video watermarking technique. This research uses a Contourlet Transform for video watermarking system, the watermark image was included in the Contourlet Transform low-frequency coefficients for each video key frames to get the best possible quality and strength against various types of video attacks. Furthermore, the performance of the method was estimated through usage infusibility analysis measurements such as the Peak Signal-to-Noise ratio (PSNR), the Normalized correlation (NC), and the Structural Similarity Index Measurement (SSIM) between original video and the watermarked one. The average PSNR,NC,and SSIM for watermarking were 55.00 dB,1.00,and 0.99, respectively, according to the test  results. Hence, Normalized cross-correlation (NCC) and correct bitrate ratio (BCR) were calculated as measures of strength between the extracted watermark and the original watermark. The results showed how this approach is very robust. It also put the system's robustness and imperceptibility to the test against various types of video attacks.