Shi Yunqing 教授学术讲座

主题：Statistical Model for Digital Image Forensics（数字图像取证的统计模型）

时间：六月十一日 （星期三），下午 三点

地点：存中楼808学术报告厅

报告人简介：Dr. Yun Qing Shi has joined Dept. of ECE at New Jersey Institute of technology (NJIT) since 1987, and is currently a professor there. He obtained his B.S. degree and M.S. degree from the Shanghai Jiao Tong University, Shanghai, China; his M.S. and Ph.D. degrees from the University of Pittsburgh, PA, USA. His research interests include visual signal processing and communications, multimedia data hiding and security, applications of digital image processing, computer vision and pattern recognition to industrial automation and biomedical engineering, theory of multidimensional systems and signal processing. Some of his research projects have been supported by some federal and New Jersey State funding agencies. He is an author/coauthor of more than 200 papers, a book and 4 book chapters. He holds three US patents, and has 24 US patents pending (20 of these pending patents have been licensed to third party by NJIT). He is the chair of Technical Program Committee of IEEE International Conference on Multimedia and Expo 2007 (ICME07), a co-chair of Technical Program Committee of International Workshop on Digital Watermarking 2007 (IWDW07), and a fellow of IEEE.

内容简介：In this talk, the role of statistical model played in digital image forensics is addressed. First, its role played in image steganalysis is described. Some typical examples are presented. Then, the talk moves from image steganalysis to image splicing (a fundamental form of tampering) detection. The difference and relationship between steganalysis and splicing detection is discussed. Along this line, a recently developed advanced statistical model of natural images learned from steganalysis and its successful performance in splicing detection are presented. The often confused term “model” is further discussed by pointing out the difference between the image model used for image compression and the statistical model developed for image forensics. Finally, statistical models used for detecting JPEG (double) compression are briefly introduced.