Research

My main research interests:

computational stereo vision (some results on real-time stereo are presented on the Photogrammetry Lab web page)
probabilistic image modelling
medical image analysis (in collaboration with Prof. A. A. Farag and Dr. A. El-Baz)
low-level image processing
3D scene understanding
space and aerial image analysis

Our eyes solve visual problems so easily that one may be puzzled why even most successful computer vision solutions rank far below in quality and processing rate. For me, human vision is a natural miracle,and more than 50 years spent in various image processing and computer vision domains suggest that it is extremely hard (if at all possible) to bridge a gap between human and computer visual skills... But, this is still a good challenge to mimic one or another side of human vision with computational techniques, especially, because there always exist particular applied problems where computer vision can amplify or replace human vision.

Every such problem is quite complex but just due to this reason is VERY attractive!
Look through some results obtained in various computer vision and image analysis areas:

computational binocular stereo,
texture modelling using Markov - Gibbs random fields,
3D scene reconstruction from calibrated multiple views, and
the least square correlation based matching of non-calibrated multiple views.

These and other results, including medical image analysis, concurrent propagation for solving ill-posed problems of global discrete optimisation, and implementation of stereo matching with massively parallel P-systems are presented in more detail in publications.