Abstract

Hough voting methods efficiently ha ndle the high complexity of multi- scale, category-level object detection in cluttered scenes. The primary weakness of this approach is however that mutually dependent local observations are in- dependently voting for intrinsically global object properties such as object scale. All the votes are added up to obtain object hypotheses. The assumption is thus that object hypotheses are a sum of independent part votes. Popular represen- tation schemes are, however, based on an overlapping sampling of semi-local image features with large spatial support (e.g. SIFT or geometric blur). Features are thus mutually dependent and we incorporate these dependences into prob- abilistic Hough voting by presenting an objective function that combines three intimately related problems: i) grouping of mutually dependent parts, ii) solving the correspondence problem conjointly for dependent parts, and iii) finding con- certed object hypotheses using extended groups rather than based on local obser- vations alone. Experiments successfully demonstrate that state-of-the-art Hough voting and even sliding windows are signi ficantly improved by utilizing part de- pendences and jointly optimizing groups, correspondences, and votes.