由網路相簿的蓬勃發展可看出網際網路的快速發展與普遍使用，但大量數位化訊息在網路上公開的流通，也引伸出數位資料被竊取後盜用的問題，已有許多數位資料的智慧財產權保護機制相繼被提出，而數位浮水印技術(digital watermarking)即是用來保護數位影像智慧財產權的一種方法。由於目前利用浮水印來做影像所有權保護的研究，最大的挑戰是浮水印的嵌入技術會影響到原圖、取出浮水印時需要原圖協助、浮水印強韌性與是否可用人眼觀點明確辨識等問題。Naor 及Shamir於1994年提出視覺密碼後，陸續有研究以結合視覺密碼與浮水印技術進行影像所有權的保護方案，許慶昇於2005年提出利用統計特性之優點來發揮視覺密碼的安全性及浮水印的強韌性，但製作出浮水印的分享影像體積較大且其方法沒有真正的運用到統計特性的精神，本研究以像素值隨機比較的方式，配合抽樣理論中的大數法則，發展出新的浮水印著作權保護方法，以期真正利用統計特性來符合視覺密碼的安全要求。 The prosperous web photo album indicates the rapid development and common usage of the Internet. However, the large amount of digital data which was widely spread on the Internet also reveals the usurpation and misappropriation problem of digital information. Many intellectual property rights protection protocols of digital information were proposed one after another. Digital watermarking is one of the ways to protect multimedia intellectual property rights. In the research of using watermark to protect image copyright, the biggest challenges are that: the inserting technique of watermarks will affect the original image. We need the assist of the original image to get the watermarks, and whether the watermarks are clear enough for people to identify it. After Naor and Shamir presented visual cryptography, research that combines visual cryptography and watermarking technique are used to protect the copyright of the image. In 2005, Hsu used the advantage of statistic feature to bring the security of cryptography and the robustness of watermark into full play. However, the size of the sharing images is four times larger than the watermarks, and the method doesn''t really use the spirit of the statistic feature. Hence, this paper use pixel comparing method as well as Large Number Theorem of Sampling Distribution, to develop a new technique for watermarking copyright protection which satisfies the safety demand of visual cryptography.
