Trainable visual models for object class recognition
Andrew Zisserman
In: Pascal Pattern Recognition and Machine Learning in Computer Vision Workshop, 3 - 5 May 2004, Grenoble, France.

Abstract

A review of the state of the art on trainable visual models for object class recognition. The tutorial covers: (i) Models that learn parts, then add structure: Weber, Welling & Perona, Leibe & Schiele, Agarwal & Roth, Borenstein & Ullman (ii) Models for which the structure is primary: Felzenszwalb & Huttenlocher, Ramanan & Forsyth (iii) Models that learn parts and structure simultaneously Fergus, Perona & Zisserman (iv) Summary and open challenges: Pascal Challenge: 101 Visual Object Classes

PDF - Requires Adobe Acrobat Reader or other PDF viewer.

[Edit]