Editorial

The five papers included here were initially presented at the 2006 conference on computer vision and pattern recognition (CVPR). The papers selected were among those considered for the best paper prize, and represent a subset of the prize committee’s view of the most valuable contributions presented at the conference. The conference “best paper” prize was awarded to Derek Hoiem, Alexei Efros and Martial Hebert for “Putting Objects in Perspective”. This paper shows how cues to 3D scene geometry can influence and improve object recognition. This has essentially always been a goal of scene understanding systems, for example Hanson and Riseman’s VISIONS (Hanson and Riseman 1987), but has proven very hard to achieve. This is the first practical system which proves the concept, and is thus an important step towards general-purpose vision systems. An honourable mention for best paper was awarded to “Learning an Alphabet of Shape and Appearance for MultiClass Object Detection” by Andreas Opelt, Axel Pinz and Andrew Zisserman. The authors show that sublinear growth in alphabet size, as had been recently observed (Torralba