|
The probabilistic vision group is part of the McGill Centre for Intelligent machines.
The focus of our research
is on developing probabilistic inference techniques for computer vision problems. Such problems often involve
processing noisy data; probabilistic approaches are then appropriate as they allow for uncertainty
to be modeled and propagated through the solution process. Choosing to represent the solution to a computer vision
problem as a probability distribution over many possible solutions makes it possible to measure the ambiguity of the results and
provide guidelines as to whether more data should be acquired, when possible. Applications include, but are not limited to
Object recognition
Active vision
Image segmentation
Image registration
Image-guided neurosurgery
|