Magic and the challenge of computer vision 
(Lectures)
6 June 2012
5.30pm
Venue: Conference Centre, 22 Symonds St, The University of Auckland
Inaugural lecture by Professor Georgy Gimel’farb, Department of Computer Science
Visual perception exploits up to 50% of the human brain for extracting about 80% of the information about the real world from a sequence of images captured by the eyes. Although the basics of human vision are already known, the whole process of perceiving and understanding the images still eludes formal explanation.
This is why computer vision, which intends to mimic in part this remarkable process, is an extremely fascinating and challenging research area. Intricate relationships between images and scene descriptions are difficult for accurate modelling. Because just the same images may depict different scenes, most of the problems are inherently ill posed and should be properly regularised to move toward a goal solution. Algorithms for the exact solutions are typically NP-hard, so that provably good, but computationally feasible approximations need to be found.
Some ways to overcome these difficulties will be exemplified in this talk by computational stereovision and medical image analysis problems.
All are welcome to this public lecture.
