ECCV
2010
Springer
14 years 5 months ago
2010
Springer
Abstract. Most current approaches to recognition aim to be scaleinvariant. However, the cues available for recognizing a 300 pixel tall object are qualitatively different from tho...