具高容量的向量地圖資訊偽裝演算法

Abstract

本篇論文係針對一新興的圖形格式：向量地圖，提出一個高容量資訊偽裝演算法。我們模擬病毒感染擴散的方法快速地決定嵌入順序，並透過QIM概念的延伸於轉換域嵌入秘密訊息，達到高隱藏量的同時仍能保有良好的視覺效果。實驗結果顯示：我們提出的演算法不需原始向量地圖與記錄額外資訊，即可達到近似於向量地圖頂點個數二倍的高資訊嵌入量及低變形程度的目的。此外，我們的方法亦可有效抵抗平移、旋轉、等向縮放及頂點重排攻擊。證實此法可適用於向量地圖資訊偽裝之領域。

We propose a high capacity steganography algorithm for the vector map. We simulate the contagious diffusion method to rapidly decide the embedded order, and through the advancing of the QIM method in the transfer domain, embed secret messages. It thereby results in high hidden capacity yet retaining well visual effects. The results indicate: Our proposed method doesn’t need the original vector map and extra recorded data during the extraction procedure. It rapidity embeds messages about twice the amount of the vertices of vector maps and still preserves well visualization of a vector map. Furthermore our method can also work effectively against translation, rotation, uniform scale and vertices reorder. This proves that this method is applicable to the field of vector maps and steganography algorithm.