本篇論文在研究感測器雜訊和相機反應函式在數位影像鑑識上的應用，我們將會對這兩個在成像過程中產生的特性分別加以討論。在感測器雜訊的部份，根據前人擷取雜訊的方法，我們提出了另一種機率空間中的特徵的表示法，我們並且用這個表示法，應用在影像來源的辨識上，和前人的作法比起來有得到比較好的辨識效果，而且透過實驗，我們得知即便是相同型號的相機所拍的照片，或是在比較低解析度的設定下所拍攝的照片，每台相機所拍攝出來的感測器雜訊仍然是穩定的，而且不同相機間的感測器雜訊是不相關的。在相機反應函式的部份，我們利用實驗的方式來驗證不同廠牌的相機會有不同的相機反應函式，並且使用單一相片擷取相機反應函式的反函式的方法和利用數位影像在經過相機反應函式之後會在影像輪廓和輪廓周圍在RGB上的顏色會呈現非線性的特性，我們並且利用平均的線性回復程度來當作特徵，並且運用這個特徵來運用在影像來源的辨識上。

This thesis studies image forensics based on sensor pattern noise and camera response function features from imaging process. We discuss these two features separately. For sensor pattern noise, we present a new feature representation in the probability space which is modeled by a previous work. We also adopt this feature representation in an application of camera source identification to achieve much satisfying result comparing with the previous work. Through the experiments, we know the sensor noise pattern is distinguishable even cameras from the same model or images with low resolution. For camera response function, we experimentally show the camera response function can be a biometric evidence of a camera. Based on the nonlinearity around the edge colors in RGB domain, we also use a linearity measurement as a feature and show an application on camera source identification based on the proposed feature.

[1] M. Kharrazi, H. T. Sencar, and N. Memon, “Blind Source Camera Identification,” Proceedings of ICIP, 2004.
[2] J. Lukas and J. Fridrich, “Digiral Camera Identification From Sensor Pattern Noise,” IEEE Trans. Information Forensics and Security, vol. 1, no. 2, June 2006, pp. 205-214.
[3] M.K.Mihcak,I.Kozintsev,andK.Ramchandran, “Spatially adaptive statistical modeling of wavelet image coefficients and its application to denoising,” Proc. IEEE Int. Conf. Acoustics, Speech, Signal Processing, Mar. 1999, vol. 6, pp. 3253-3256.
[4] J. Armoando Dominguez-Molina, G.. Gonzalez-Farias, and R. M. Rodriguez-Dagnino, “A practical procedure to estimate the shape prarmeter in the generalized Gaussian distribution,” CIMAT Tech. Rep. I-01-18_eng.pdf. Available: http://www.cimat.mc/re-portes/enlinea/I-01-18_eng.pdf.
[5] S. Lin, J. Gu, S. Yamazaki, and H.-Y. Shum. “Radiometric Calibration from a Single Image,” Proceedings of CVPR, 2004, pp. 938-945.
[6] S. Lin and L. Zhang, “Determining the radiometric response function from a single grayscale image,” Proceedings of CVPR, June 2005, vol. 2, pp. 66-73.
[7] T.-T. Ng and S.-F. Chang, “Camera response function estimation from a single grayscale image using differential invaraints,” Tech. Rep., ADVENT, Columbia University, 2006.
[8] F. M. Candocia, “A semiparametric model for accurate camera response function modeling and exposure estimation from comparametric data,” IEEE Trans. Image Process., vol. 14, no. 8, August 2005, pp. 1138-1150.

[9] S. Mann, “Comparametric equations with practical applications in quantigraphic image processing,” IEEE Trans. Image Process., vol. 9, no. 8, pp. 480-487.
[10] T. Mitsunaga and S. K. Nayar. “Radiometric self calibration,” Proceedings of CVPR., 1999, pp. 374-380.
[11] M. D. Grossberg and S. K. Nayar. “What is the space of camera response functions?” Proceedings of CVPR, 2003, pp. 602-609.
[12] Z. Lin, R. Wang, X. Tang, and H-Y Shum. “Detecting doctored images using camera response normality and consistency,” Proceedings of CVPR, 2005, vol. 1, pp. 1087-1092.
[13] Y-F. Hsu and S-F. Chang, “Detecting Image Splicing using Geometry Invariants and Camera Characteristics Consistency,” Proceedings of ICME, 2006, pp. 549-552.
[14] C-C. Chang and C-J. Lin, LIBSVM : a library for support vector machines, 2001. Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm.
[15] dcraw available at http://www.lebsanft.org/blog/index.php?cat=9.