| Summary: | 碩士 === 國立清華大學 === 電機工程學系 === 94 === As the internet and storage equipment become more and more popular, the digital data can be distributed easier and faster. The pirates can easily and unlimitedly copy the digital multimedia (images,audio, and video). Therefore, people are more and more attentive to the copyright protection for digital multimedia.
Digital watermarking is a useful and developing technology, it is the direct embedding of additional information into the original content or host signal. Ideally, there should be imperceptible difference between the watermarked and original signal, and the watermark should be difficult to remove or alter without damaging
the host signal. In some instances, the amount of information that can be hidden and detected reliably is important. Applications include copyright protection, owner identification, and broadcast monitoring. The evident difference between the watermarking and encryption is that the data is only protected during the transmission using the traditional cryptography. After decrypting, the digital data is no longer protected, it can still be copied and
distributed. In contrast to traditional cryptography, anyone can use the multimedia data, once the dispute about the ownership of data happened, we can extract the information from the host data to authenticate and characterize the owner by watermarking technique.
A lot of researches discussed how to embed a watermark into image in recent years. For digital watermarking of image, it often embedded in the spatial/frequency domain. Watermark can be embedded into the image intensity in spatial domain or into image coefficients in transformed domain. In spatial domain, watermark is usually embedded
into low bits to make it imperceptible. In transformed domain, the image is first orthogonally transformed, then watermark is embedded into transformed coefficients. However, few researches analyzed the watermark techniques in video. As far as the watermarking of a video sequence is concerned, in the literature two approaches are
considered: watermark insertion on the original raw video data or alternatively on the coded data bit stream such as motion vector. Digital content often suffer some intentional or unintentional attacks such as compression, cropping, add noise, and geometric attack, etc. Therefore, we need to consider those of attacks for
designing the watermarking system.
In this thesis, we propose a watermarking system based on block matching algorithm. The advantages of our proposed system are not only the robustness in the signal attack and invisibility but also the robustness in the intentional attack such as cropping. The watermark is mainly embedded in the uncompressed domain, and it is detected without the use of the original video information. Hence, this scheme is powerful and possesses the adjustable threshold value
to obtain the balance relationships between the imperceptibility and robustness. For the security viewpoint, the watermark embedding usually requires the knowledge of a secret embedding key to perform. The experimental results show that our proposed method can
successfully detect the watermark from the attacked video especially for video compression.
|
|---|
