Generic object detection is a very interesting problem that can find
uses in a large variety of applications from image context search, to
robotic navigation.  The recent PASCAL Visual Object Class (VOC)
competition has provided a test bed to compare may of the current
cutting edge techniques.  This talk will discuss general methodologies
for object class presence detection, and localization.  We will focus
primarily on those techniques that proved successful in the PASCAL VOC
challenge, as well as try to bring to light some of the difficulties
that all of these systems face.  We will also discuss the techniques we are
currently actively developing, as well as introduce our generic
framework for object class localization.