Goal: To contribute to
solutions of
recognition problems, relevant to the human
visual system and to human comfort.
We have been
addressing the
following issues:
Orientation and
frequency description of still and video images for general
purposes,
Decompositions based on Gabor
filters,
Dense orientation maps,
Interpretation and modelling of dense orientation maps via differential
operators (e.g. tensors, Lie groups), modelling dense symmetry
fields.
We have been using
our results in the following applications:
Biometric person
authentication by audio and video, face recognition, fingerprint
recognition
Content based multi-media data base queries
Localization, identification, and tracking of tagged objects in video;
Unsupervised image segmentation, and texture analysis.
Projects (Select):
-
Network of excellence: BIOSECURE (EU research school on biometric identification)
- OCR for modification based
scripts: Ethiopian (Swedish SIDA )
- OCR for modification
based scripts: Sinhala (Swedish SIDA)
- Image
content query and retrieval for proteomics (Swedish SSF)
Members :
Former Members:
Downloads
|