Colloquium "Object recognition - overview, perspectives and challenges"

Abstract:

Object recognition is a field of computer vision that is concerned with building models of visual concepts. In the first part of the talk a short overview of methods for object and scene categorization was given, with an emphasis of bag-of-words methods that are over a decade a state-of-art image representation for object and scene categorization. In the second part the method for classification and detection of object instances, like products, landmarks and logos was briefly described. In the last part an overview of current trends in object recognition, outstanding challenges and perspectives was outlined.

CV:

Josip Krapac has finished undergraduate and master studies in computer science on Faculty of Electrical Engineering and Computing, Zagreb, Croatia. During his PhD, shared between GREYC group of University of Caen and LEAR group in INRIA Grenoble, he worked on image representations for classification and ranking. After graduation he did a post-doc in TexMex team in INRIA Rennes concerning logo and product classification. His research interests concern the representations of images, application of machine learning methods for solving the problems in vision, in particular for object categorization and localization.