




By: Guangsen Tian
Advisor: Professor Yun Q. Shi
Department of Electrical and Computer Engineering
Time: 12:00-1:00 PM, Wednesday, May 4th, 2005
Place: Room 202, Kupfrian Hall, New Jersey Institute of Technology (NJIT), Newark NJ. Directions
Abstract
Recently, a novel reversible data embedding method is reported, which is based on difference expansion (DE). By exploring redundancy in digital images it achieves high embedding capacity, while keeping visual distortion of stego-image low. In this thesis, this technique is firstly studied and experimentally evaluated. Secondly, an effective steganalysis scheme for this DE-based reversible data embedding method is proposed, which uses 12-D feature vectors and Bayes Classifier. The proposed steganalysis scheme has steadily achieved a correct classification rate of 99%.
Committee Members:
Associate Professor Constantine N. Manikopoulos
Professor Yun Q. Shi (chair)
Professor Mengchu Zhou



