|
Feature-based image analysisMonday January 14 at 2.00 pm Martin Lillholm will defend his halfway research report according to the 4 year PhD regulations in room 2.63 at the IT-University, Glentevej 67, DK-2400 NV.AbstractAccording to Marr's paradigm of computational vision the first process is an extraction of relevant features. A stream of artificial vision research has followed this approach. The aim of this presentation is to quantify and characterize the information on the image that features carry and in this way indirectly evaluate this concept of feature-based image analysis. The main conclusion is that a reasonably low number of features characterize the image to such a high degree, that nearly all information in the image is revealed. The strategy is to define metameric classes of images and examine the information contents in a canonical least informative representative of this class. Algorithms for identifying these are given. Finally, optimal feature detectors localizing the most informative points in images are derived on basis of definitions of statistics of images.After the presentation and a short break Martin will be examined by Ass. Prof. Luc Florack, Technical University of Eindhoven |