Thesis (Ph.D.) - Cairo University - Faculty of Computers and Artificial Intelligence - Department of Information Technology

Over the past recent years, digital watermarking technology has seen much use and has been viewed as a potential and effective method of ending illegal access to digital multimedia products. Visual watermarking programs enables the protection of digital videos and images, which are produced based on specific demands, like contents intended for digital library or materials for distant learning websites. These contents prohibit unlawful reproduction or copying of the material. Visible watermark mechanism defends IPR (Intellectual Property Rights) persuasively; openly, watermarked data mainly contains discoverable but un-destroyable copyright keys showing the identity of IPR owners. Many commercial TV stations use video logos, which is a form of visible watermark indicating a disclaimer of content ownership. The video logos also act as a symbol of authentication to re-stream the videos when the primary logos are used together with the new logos .A regretful side effect of these logos is the constant reduction in viewing joy. Until there can be complete removal of the watermark patterns without the destruction of the video quality of the protected content, people cannot directly use visible watermarked information. Meanwhile, abundant studies are discussing the security of invisible watermarks, but there are microscopic researches on the significance of visible watermark technology Our research objective is divided into two sub-problems. The first sub-problem is to locate the visible watermark on video frames with high accuracy in order to be removed. and the second sub-problem is to consider the watermarked area/s as a damaged area/s to be reconstructed exploiting spatial and temporal information

Issued also as CD