Font Size: a A A

The Studies For Video Watermarking Algorithms Based On Features

Posted on:2018-07-19Degree:MasterType:Thesis
Country:ChinaCandidate:S J YuanFull Text:PDF
GTID:2348330515966754Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
With the rapid development of multimedia technology,the video plays a more and more important role in our life,work and study.In recent years,the cases for video piracy and content tampering are increasing.It is urgent to protect the video copyright.To solve the problem of video copyright protection,the video watermarking algorithms with effectiveness,invisibility and robustness have been proposed.However,there are only a few video watermarking algorithms based on features.In this thesis,two video watermarking algorithms based on features for copyright protection are proposed.1.The video watermarking algorithm based on region selection is proposed to balance the robustness and invisibility.Firstly,an improved structure tensor operator is used to select the best 4 × 4 block as the embedding block of watermark information.Secondly,the luminance channel of I frame is chosen to embed the watermark information.The selected 4 × 4 pixel block is operated using DCT(Discrete Cosine Transform)and SVD(Singular value decomposition).Finally,the watermark information is embedded into the video by the odd-even rule in the algorithm.Experimental results show that the watermarking algorithm can protect the video copyright well.The PSNR of each frame is greater than 45,and the SSIM is at least 0.99.At the same time,the watermark extraction rate is more than 90% after some common attacks,such as compression,noise,filtering,and frame dropping,frame adding.2.An adaptive video watermarking algorithm is proposed with different embedding intensity according to different texture and brightness in the video.Firstly,each 16 × 16 macro block is transformed into two levels of DWT(Discrete Wavelet Transform).Secondly,the low frequency components of DWT are transformed by DCT to reduce the impact of embedding process.Finally,the watermark information is embedded into the video by the calculated adaptive quantization step size and odd-even rule.Adaptive embedding quantization step is calculated for macro block based on Radon transform.Experimental results show that our watermarking algorithm is invisible and robust against common attacks.
Keywords/Search Tags:Video Watermarking, Video Features, Region Selection, Adaptive Quantization Step, Radon Transform
PDF Full Text Request
Related items