Digital Image Watermarking Based on Fractal Compression Scheme

[[abstract]]二十世紀末的數位革命，為人們帶來了無限的可能與便利，網際網路取代了傳統的傳播媒體成為最有效的傳媒工具。再加上多媒體的發展，使得任何最新的創作財產都能以豐富多變的面貌傳遞給每一個人。這時智慧財產權的問題就浮現出來，因為每個創作者和版權所有者都不希望自己的作品被剽竊或複製。因此為了保護每個合法擁有者的權利，數位浮水印的技術逐漸受到重視，也吸引更多的學者投入研究浮水印的技術。 本碩士論文是利用碎形原理來發展新的浮水印隱藏法；傳統碎形的浮水印技術是藉由碎形編碼的搜尋過程，將浮水印資訊隱藏在碎形碼中的位置參數或旋轉參數中。本論文利用次佳化最小平方誤差逼近法（sub-optimal least squares approach）來調整碎形碼中的灰階轉換參數，將浮水印資訊隱藏在碎形碼中。相較於傳統的方法，我們所提出的演算法在抵抗JPEG壓縮的攻擊方面有較佳的效果，並可抵抗覆蓋（overlapping）、雜訊干擾及塗鴉等各種攻擊。[[abstract]]In the end of 20’th century, the digital revolution brings limitless convenience for people. Internet has taken the place of conventional medium. It becomes the most efficient medium. The creations show people rich and colorful appearance by the Internet and multimedia. The protection of intellectual property is very important. The creator and copyright owner don’t wish their work would be cribbed. The digital watermarking technology has been respected in order to protect every lawful owner’s rights. And more investigators are devoted to the study of this issue. In this thesis, our algorithm is a new way based on fractal compression scheme. The previous technology embeds the watermark into position’s parameters and rotation type of fractal code with searching procedure in fractal compression. We adjust gray mapping transforms parameters by way of sub-optimal least squares approach in order to embed the watermark into the fractal code. Compared with the previous skills, the algorithm has better result in resisting JPEG compression. And it could resist the overlapping, daubing, noise and various attacks