A Robust Discrete Wavelet Transform Based Adaptive Watermarking Scheme in YCbCr Color Space against Camcorder Recording in Cinema/Movie Theates

Abstract

Camcorder-based piracy is increasing nowadays because of the high quality of camcorder devices. The most affected industries are theatre and cinema. To defeat theatre piracy which causes huge loss to the motion picture industry of the world, here in this paper, a robust discrete wavelet transform (DWT) based video watermarking approach has been presented in which a payload formed with date-time of the show, registration number of theatres, and six-digit zonal Improvement code (ZIP) of the theatre location embedded in the movie before it plays back in the theatre. To assure that the video and the algorithm are of good quality and robustness, the adaptive watermarking pattern has been used to minimize the watermarking area in the host video. In this study, YCbCr color space has been adopted to use luminance factor of the image to watermark because the luminance part of the image is the least sensitive to the human visual system. This approach shows the convincing robustness against the geometrical attack and camcording attack, which is the most affecting factor in camcorder piracy. As the proposed approach is good enough to disclose the place and time of piracy, it may deter the pirate to do piracy.

Spatio-Temporal Context Based Adaptive Camcorder Recording Watermarking

“…Therefore, in Section 3.1, we propose to embed the watermark by constructing a temporal histogram. In terms of watermark embedding, [25] and [11] select ixed small blocks to modify their pixel values, while we construct a spatial histogram based on content features to improve the video quality. It should be noted that 1) Adjacent frames are diferent in changing scenes or high motion scenes.…”

“…For synchronization problems caused by frame rate changes, we can ix the frame rate during watermark embedding and extraction. Regarding the problem of watermark synchronization, literatures [25], [1], [28] and [11] use a synchronization sequence to locate the watermark, but this is a non-blind extraction that requires original information. So we put forward a watermark correlation-based synchronization algorithm and combine it with cross-validation in Section 3.2.1, which can resist temporal sampling attack and does not require additional information.…”

“…These distortions may destroy the watermark by modifying the pixel values. In order to enhance the watermark robustness, similar to [25] and [11], we modify the blocks of the whole frame to embed the watermark as shown in Fig. 3.…”

“…Identify whether the video contains watermarks or not. The traditional methods embed a synchronous sequence in the video [1,11,25,28,34]. If the synchronous sequence is extracted, the video is considered to be watermarked.…”

“…However, these methods do not consider the video content characteristics, so there will be video licker in the visual efect. To decrease the modiied area and video licker, the researchers adjust the luminance of a ixed number of small blocks to embed the watermark and deine the embedding strength based on the local contents [11,25]. However, the positions of these modiied blocks are ixed, which may cause visible licker to the human eyes in some local areas.…”