植基於紋理特徵之數位浮水印技術於數位影像竄改驗證與回復的應用

由於數位訊息技術的快速發展，網路和數位消費平台已經深刻的影響了我們的日常生活以及社會型態，這樣的情形導致大量的數位信息，在傳送以及儲存時變得非常方便，然而有關資料安全性方面的議題，在這便需要謹慎的去面對。例如，它們變得容易被修改、複製、及傳送分發，因此，有關於防範方法被相繼的提出，常見的數位版權保護技術如下，數位認證技術以及資訊隱藏等技術。雖然已有許多的防護技術存在，但如何做到更好的防護以及認證效果，仍然是一個核心的重要課題。本論文提出兩種碎型浮水印技術，分別在頻率域及空間域中提取特徵浮水印。首先，提出的第一個方法為一種新型的碎型浮水印，藉此來達到影像內容的認證、竄改定位和影像的恢復，所提出的方法是針對受保護的影像，採用離散小波轉換出近似的子頻帶，再對此頻帶用一種新型的區域二元化型態方法取出特徵浮水印，進而根據三種類型的浮水印，去呈現出三種不同的圖像特徵，藉此來實現的多層的浮水印認證，除此之外，特徵浮水印也包含恢復圖像的資訊。該方案在不同的攻擊中，可以有效的實現竄改檢測，如在複製攻擊和拼貼攻擊中，也能夠保障其安全的身份驗證，不僅如此，它還具有相當程度的回復效果。本論文所提出的第二個方法是另一種新型的碎型浮水印技術，主要使用LDMP以及LBP的方法來產生特徵浮水印。這些浮水印是從空間域中取出，它具有更精確的驗證功能，其驗證的最小範圍為圖像中的每一個像素點，同時也具有影像回復之機制。然而為了提高浮水印的安全性，我們藉由引用混沌映射的方法來達成此目的，它可以擾亂像素以及浮水印之間的嵌入位置，同時又能保有兩者之間的初始關係，能使浮水印安全性大幅提升。

關鍵字

碎型浮水印；離散小波變換；區域二元化型態；多層式浮水印；混沌映射；竄改檢測；圖像認證及恢復

並列摘要

Due to the rapid development of digital messaging technology, the Internet and digital consumer devices have profound impact on our daily life and social habits. For this reason, a large number of digital messages become very convenient for transporting and storing. However, this raises a serious problem about data security. For example, the digital ones are very easy to be modified, duplicated, and distributed. Therefore, many protection schemes were proposed, common digital copyright protection are digital authentication technology and data hiding. Today, many protective technologies are proposed but how to achieve a better protective effect. It is still an important issue. This thesis presents two fragile watermarking schemes individually in the spatial domain and transform domain according to the image signification features. The first proposed scheme is a novel fragile watermarking scheme for image content authentication with tampered localization and image recovery. The proposed scheme uses a novel Local Binary Pattern (LBP) method to generate feature watermark from the LL sub-band of the discrete wavelet transform (DWT), then it constructs 3-types feature watermark to present three different types of image characteristic to achieve the multilayer authentication. Moreover, the feature watermark contains recovery information. The proposed scheme not only can achieve tamped detection well under different attacks, such as copy-and-paste attack and collage attack, but also can safeguard the secure authentication. Moreover, it also has good recovery capability. The second proposed scheme is another novel feature watermarking scheme. The proposed method use Local Different Mean Pattern (LDMP) and Local Binary Pattern (LBP) to generate feature watermark from original image. Moreover, the feature watermark has the ability to detect tampered blocks and recovery error blocks. The minimum detection range is the pixel size. In order to improve the security of the proposed scheme, a chaotic map is employed. It can maintain the relationship between pixels in the watermarked image and the disturbed one, and helps the watermarking scheme to resist impersonation attacks.

並列關鍵字

Semi-fragile watermarking ； Discrete wavelet transform ； Local binary patterns ； Multi-level watermark ； Chaotic map ； Tampered detection ； Image authentication and recovery

參考文獻

[1] C. Cruz, R. Reyes, M. Nakano and H. Perez, “Image content authentication system based on semi-fragile watermarking,” in Proceedings of the 51st International Conference of Midwest Symposium on Circuits and Systems, Mexico, 2008, pp.306-309.

[3] D. Shoba, M. Karabacak, S. Periandy and S. Ramalingam, “FT-IR and FT-Raman vibrational analysis, HF and DFT simulations of isocyanic acid 1-naphthyl ester: Molecular and Biomolecular Spectroscopy,”International Journal of Spectrochimica Acta Part A, Vol. 1, No. 15, 2008, pp. 504-518.

[6] H. Yang and X. Sun, “Semi-fragile watermarking for image authentication and tamper detection using HVS model,” in Proceedings of International Conference on Multimedia and Ubiquitous Engineering, USA, 2007, pp. 1112–1117.

[7] I.K. Yeo and H.J. Kim, “A novel audio watermarking scheme,” in Proceedings of International Conference of information Technology: Coding Computing, 2001, pp. 237.

[10] N. Chotikakamthorn, “Document image data hiding technique using character spacing width sequence coding,”International Conference of IEEE on Image Processing, Kobe, Vol. 2, 1999, pp. 250-254.

國際替代計量

植基於紋理特徵之數位浮水印技術於數位影像竄改驗證與回復的應用

全文下載

主題瀏覽