供資訊隱藏應用之大量資料植入與搜尋技術之研究

Abstract

隨著電腦科技的進步與網路的普及，越來越多的資料可在網路上快速且便利的傳送。本論文首先提出在GIF檔案上的一個大量且無損的資訊隱藏法。在這方法中，我們複製影像調色盤中出現頻率高的顏色到調色盤中沒用到的空格中，並且利用被複製的顏色來藏入資訊。實驗發現，用出現頻率越高的顏色來藏入越多位元，整體隱藏量會提高許多。接著，我們提出在PNG影像上一個可以在網頁上植入祕密資訊的方法。此方法的總隱藏量取決於前景色PNG的透明度通道所使用的隱藏位元數。隱藏完之後，我們利用提出的方法來調整前景圖和背景圖的灰階強度來減低資訊隱藏後所造成的雜訊。最後，我們提出兩個可在BMP影像資料庫中快速找尋所需影像的方法。一個是非區塊層次(non-block-level)的方法，在此方法中我們把註解資訊循序的藏到影像裡; 另一個則是區塊層次(block-level)的方法，在此方法中我們先把影像分成區塊後並分別把註解資訊藏到區塊裡。實驗結果顯現出這些方法的靈活度與實用性

With the advance of computer technologies and the popularity of the Internet, more and more data can be transmitted speedily and conveniently on public networks. In this study, first we propose a distortion-free and high-capacity data hiding method on GIF. This method duplicates colors of high frequencies to fill the unused color entries in the palette and uses the duplicated colors to hide secret message bits. It is found that more bits can be hidden by the method if colors with higher appearance frequencies are duplicated first. For PNG images, we propose a data hiding method for implanting secret messages in web pages. The hiding capability is achieved by changing the transparency values of the alpha channel of the pixels of the foreground PNG image. A scheme to adjust the intensity values of both foreground and background image pixels to reduce the artifacts caused by data hiding is also proposed. Finally, we propose two fast methods for searching desired BMP images in databases. One is a non-block-level method that hides comments in images sequentially, while the other is a block-level method that divides images into blocks in which comments are embedded. Good experimental results show the feasibility and applicability of the proposed methods.