半脆弱影像浮水印通常能夠抵抗如影像壓縮等非惡意的攻擊，進而達到認證保護的目標。然而為了改善浮水印的強健度，本篇論文提出了基於小波空間的影像特徵之半脆弱影像浮水印的演算法。因為在小波子頻帶的係數存在著以統計方法求得的關係，基於以上的理由，在此篇論文中我們使用的是在三個小波子頻帶中，相對應的資料區塊所求得的變異數。求得變異數的方法則為高斯混合模型(Gaussian Mixture Model, GMM)以及最大期望值演算法(Expectation Maximization Algorithm, EM)。比較在三個小波子頻帶中於相同位置的區塊所計算出來的變異數(Variance)，根據欲嵌入影像的浮水印資訊來調整三個變異數之間的關係。然而，當影像經過旋轉的處理之後，在影像方形區塊中的像數值將會變得不同。因此，為了抵抗旋轉的攻擊，將子頻帶中每一個影像區塊以同心圓的方式作切割，而且在本篇論文中，這些便是我們要處理的區塊。如果需要藏入更多的浮水印資訊時，可以更進一步的將這些同心圓的環狀區塊切割成塊狀區塊。當然，再切割成塊狀區塊之前，必須先找出一個不變的方向作為切割的根據以便能夠正確的找回切割的區塊。因此需要先使用Scale Invariant Features Transform(SIFT)演算法求得在原始影像中的特徵點，而具有最大強度的特徵點則被當成同心圓中將環狀區塊切割成塊狀區塊的起始角度。因為此具有最大強度的特徵點在經過旋轉的攻擊之後是不會改變而且可以找到的，所以可以找回當初藏入的浮水印資訊。根據實驗的結果顯示，本篇論文所提出來的浮水印演算法不僅可以抵抗旋轉的攻擊，效能也優於其他的演算法。

Semi-fragile image watermarking is usually to achieve authentication protection against some non-malicious attacks such as image compression. In order to improve the robustness of watermarks, a feature-based semi-fragile watermarking in wavelet domain is proposed in this paper. Because some statistic relations between the coefficients in wavelet sub-bands are existed, the variances of the same regions locating in each sub-band are used in this paper. The variance of a region are generated by expectation maximization (EM) algorithms and Gaussian mixture model (GMM), and the variances of the regions with same position in the three high frequency bands are compared and adjusted according to the watermarks embedded. However, the pixels of a square region in an image are different while the image is rotated. For resisting the rotation attack, every sub-band is divided into several concentric circles, and the circle regions are processed in this paper. The circle regions can also be divided into several slices if higher capacity of information hiding is required. However, it is necessary to find the invariant direction of an image for seeking the correct slices in a circle region. Therefore, the feature points in the original image are detected by scale-invariant feature transform (SIFT), and the direction with the maximum magnitude of feature point is regarded as the start line of the circle regions. Because the direction is detectable and invariant, the watermarks embedded in the same slice can still be extracted correctly after rotation attack. According to the experimental results, the proposed watermarking algorithm not only can resist the rotation attacks but also has better performance than others.

[1] Central Bank of the Republic China (Taiwan), http://www.cbc.gov.tw/
[2] C. Harris and M. Stephens, “A combined corner and edge detector,” Fourth Alvey Vision Conference, 1988.
[3] C. Rey, and J. L. Dugelay, “A survey of watermarking algorithms for image authentication,” EURASIP Journal on Applied Signal Processing, 2002.
[4] C. Rey, and J. L. Dugelay, “Blind detection of malicious alterations on still images using robust watermarks,” in Secure Images and Image Authentication Colloquium, 2000.
[5] C. T. Hsu and J. L. Wu, “Hidden digital watermarks in images,” IEEE Transactions on Image Processing, vol. 8, no. 1, January 1999.
[6] C. Y. Lin, and S. F. Chang, “Semi-fragile watermarking for authenticating JPEG visual content,” SPIE Security and Watermarking of Multimedia Contents II, vol. 3971, January 2000.
[7] C. Y. Lin, and S. F. Chang, “A robust image authentication method distinguishing JPEG compression from malicious manipulation,” IEEE Transactions on Circuits and System for Video Technology, vol. 11, no. 2, February 2001.
[8] D. G. Lowe, “Distinctive image features from scale-invariant keypoints,” International Journal of Computer Vision, 60, 2, 2004.
[9] D. G. Lowe, “Object recognition from local scale-invariant features,” International Conference on Computer Vision, 1999.
[10] D. Zou, Y. Q. Shi, Z. Ni, and W. Su, “A semi-fragile lossless digital watermarking scheme based on integer wavelet transform,” IEEE Transactionson Circuits and Systems for Video technology, vol. 16, no. 10, October 2006.
[11] E. Koch and J. Zhao, “Towards robust and hidden image copyright labeling,” IEEE Workshop on Nonlinear Signal and Image Processing, January 1995.
[12] H. Yuan and X. P. Zhang, “A multiscale fragile watermark based on the Gaussian mixture model in the wavelet domain,” International Conference on Acoustics, Speech, Signal Processing, 2004.
[13] H. Yuan and X. P. Zhang, “Multiscale fragile watermarking based on the Gaussian mixture model,” IEEE Transactions on Image Processing, vol. 15, no. 10, October 2006.
[14] I. J. Cox, J Kilian, F. T. Leighton, and T. Shamoon, “Secure spread spectrum watermarking for multimedia,” IEEE Transaction on Image Processing, vol. 6, no. 12, December 1997.
[15] J. K. Romberg, H. Choi, and R. G. Baraniuk, “Bayesian tree-structured image modeling using wavelet-domain hidden Markov models,” IEEE Transactions on Image Processing, vol. 10, no. 7, July 2001.
[16] K. Maeno, Q. Sun, S. F. Chang, and M. Suto, “New semi-fragile image authentication watermarking techniques using random bias and nonuniform quantization,” IEEE Transactions on Multimedia, vol. 8, no. 1, February 2006.
[17] M. Brown and D. G. Lowe, “Invariant features from interest point groups,” British Machine Vision Conference, 2002.
[18] M. U. Celik, G. Sharma, E. Saber, and A. M. Tekalp, “Hierarchical watermarking for secure image authentication with localization,” IEEE Transaction on Image Processing, vol. 11, no. 4, April 2002.
[19] P. Bas, J. M. Chassery, and B. Macq, “Geometrically invariant watermarking using feature points, ” IEEE Transactions on Image Processing, vol. 11, no. 9, September 2002.
[20] P. W. Wong, “A public key watermark for image verification and authentication,” International Conference on Image Processing, vol. 1, 1998.
[21] Q. B. Sun and S. F. Chang, “Semi-fragile image authentication using generic wavelet domain features and ECC,” IEEE International Conference on Image Processing, September 2002.
[22] S. Pereira and T. Pun, “Robust template matching for affine resistant image watermarks,” IEEE Transactions on Image Processing, vol. 9, no. 6, June 2000.
[23] X. Wu, J. Hu, Z Gu, and J. Huang, “A secure semi-fragile watermarking for image authentication based on integer wavelet transform with parameters”, Third Australasian Information Security Workshop, January-February, 2005.
[24] Z. M. Lu, D. G. Xu, and S. H. Sun, “Multipurpose image watermarking algorithm based on multistage vector quantization,” IEEE Transactions on Image Processing, vol. 14, no. 6, June 2005.