File is GBRES.HTM
The group's research on image processing concentrates on the
extraction of structural and symbolic features from colour, range and
thermal images of 3-D objects. The symbolic features are in the form of
an attributed relational graph. This describes the relations between
contours of regions in a segmented image. Numerical attributes specify
the colour ratios, position in image coordinates, invariant combinations
of body centred moments, parameters of a best fitting plane (for
More details on image processing research Here
The group's research on computer vision is mainly directed towards
interpreting the structural features extracted by the image processing.
Graph matching is the main tool for recognition. A particular interest
as is the study of control strategies that are designed to prevent the
matching becoming caught in the NP trap. A pilot system has been
developed that exploits a coarse grained MPMD environment.
Separate research has developed a system for inspection of
foodstuffs that is based on the processing of transmitted infra-red and
the comparison of the resulting image with a structural model of the
foodstuff being inspected. The system has been applied to the inspection
of live crabs and a patent has been obtained.
More details on computer vision research
The group's research on graphics is mainly designed to support the
research on computer vision through a study of modelling techniques to
create reference objects for object recognition from CAD/CAM and
More details on graphics research Here
The group's research on virtual reality is centred on the creation
of realistic models of the ancient city of Trondheim in time for the
1000 th. year of the city in 1997.
More details on virtual reality research Here
Back to homepage Here