PhD Qualifying Examination


"Image/video tampering and its detection"

Mr. Junfeng He


Abstract:

With the advent of image/video tampering techniques such as
image/video cut out, pasting, matting, inpainting and so on,
photorealistic composite images/video can be created effectively
and conveniently. On the one hand, these techniques are important
for lots of applications, especially those in movie or entertainment
industry.  However, on the other hand, they also cause a new
problem: how to develop techniques to find out those fake
images/videos when seeing is not believing anymore. In this paper,
we review papers in both these two related topics: image/video
tampering, and detection of image/video tampering.  Briefly
speaking, image tampering techniques try to create perceptually
realistic image/videos, leaving seamless composite material
and no artifacts to human eyes; while detection of image tampering
does its job by checking whether the image/video are physically
realistic, or in particular, by checking the consistency in
physical properties of different parts.  There is a gap between
perceptually based vision and physically based vision.
Questions has long been asked and thought that: How large
is this gap? How to make use of this gap? How to cross this
gap? The study of image tampering and its detection may be helpful
to answer this fundamental but important question.


Date:     		Thursday, 24 August 2006

Time:                   10:00a.m.-12:00noon

Venue:                  Room 3311
			lifts 17-18

Committee Members:      Dr. Chi-Keung Tang (Supervisor)
			Dr. Albert Chung (Chairperson)
			Dr. Philip Fu
			Dr. Jiaya Jia (CUHK)


**** ALL are Welcome ****