Abstract
We propose a unified approach for bottom-up hierarchi-cal image segmentation and object candidate generationfor recognition, called Multiscale Combinatorial Grouping(MCG). For this purpose, we first develop a fast normal-ized cuts algorithm. We then propose a high-performancehierarchical segmenter that makes effective use of multi-scale information. Finally, we propose a grouping strategythat combines our multiscale regions into highly-accurateobject candidates by exploring efficiently their combinatorial space. We conduct extensive experiments on both the BSDS500 and on the PASCAL 2012 segmentation datasets, showing that MCG produces state-of-the-art contours, hi-erarchical regions and object candidates.