COMPUTER VISION GROUP

Our Goal is

with scientific image and signal analysis methods to contribute to the solution of recognition problems which are relevant to the human visual system and to the human comfort.

We have been addressing the following issues

Orientation and frequency decomposition of still and video images for general purposes such as image segmentation, motion tracking, and object recognition. In particular we have been working on decompositions based on Gabor filters, and their fast equivalents which are based on the symmetry descriptions of the local power spectrum.
Biometric person authentication by audio and video, face recognition
Content based multi media data base queries
Unsupervised image segmentation, and texture analysis