整合人類視覺系統感知遮罩之彩色影像浮水印技術

Abstract

近年來，許多學者致力於數位浮水印研究，以解決數位內容所有權保護與驗證之議題。浮水印依是否可被肉眼所見分為可視與不可視，本論文中，我們將可視浮水印嵌入彩色圖片數位內容之中，可視浮水印將對所有權保護產生效用，防止有心人對可視浮水印圖片進行破壞，使其失去所有權保護之效能。

在設計可視浮水印系統時，為了在半透明性與強韌性兩者間取平衡，本研究中，我們發現人類視覺系統 HVS (Human Visible System)中的對比敏感函數 CSF (Contrast Sensitive Function) 與雜訊可見函數(Noise Visibility Function) 可以達成此目標，透過CSF與NVF可以得知圖片全域與區域性的特性以便進行可視浮水印的嵌入。再加上不可視浮水印，利用Min-Jen Tsai et al. 在2008年六月提出來的半易碎浮水印驗證與回復的架構[5]，使得我們的浮水印除了可以保障數位影像的合法使用權外，同時也擁有驗證影像竄改，以及竄改後還能將影像回復的功能。

實驗結果顯示，我們所設計的可視浮水印圖片除了半透明性比較好外，抵抗非惡意的影像處理後，人眼仍能感受到良好的視覺品質，並且在影像重建理論方法下之抹除可視浮水印測試，亦能獲得較佳之抵抗結果。如果藏有浮水印的數位影像被人竄改了，亦可利用[5]提出的影像驗證浮水印技術可以偵測並標示出這個竄改區域，進而回復。

Recently, many researchers invest in developing digital watermarking technology to secure the ownership and the authenticity of the media. There are mainly two types of watermarking algorithms: visible watermarking and invisible watermarking. In this thesis, we propose a visible watermarking techniques to solve the issue of copyright protection and authentication for digital image.

In order to get the best tradeoff between the embedding energy of watermark and the perceptual translucence for visible watermark, the contrast-sensitive function (CSF) and noise visible function (NVF) of perceptual model is applied to the global and local image properties. The perceptual model identifies texture and edge regions to determine the optimal watermark strength at the watermark embedding stage. In addition, the effects of CSF masking which we propose not only fully match with the inverse of CSF, but also have better effects of contrast sensitivity than the traditional square function. The proposed method resolves the problem caused by multiple thresholds . The success is due to the fact that the function of this method is used in place of many thresholds to reduce weights in low frequency domain.

From the experimental results, the proposed technique find a searches between visibility,transparency and robustness, these are requirements that visible watermark is usually satisfied . The proposed method not only has better visibility than other algorithms, but also preserves robustness under various signal processing and advanced image recovery attacks. Visual Quality Measurement show good performance for our proposed visible watermark in VSNR,MSSIM and NQM.