Semantic contours from inverse detectors

Westudythechallengingproblemoflocalizingandclassifying category-specific object contours in real world images. For this purpose, we present a simple yet effective methodforcombininggenericobjectdetectorswithbottomup contours to identify object contours. We also provide a principledwayofcombininginformationfromdifferentpart detectors and across categories. Inorder tostudy theproblem and evaluate quantitatively our approach, we present a dataset of semantic exterior boundaries on more than 20,000 object instances belonging to 20 categories, using theimages from theVOC2011 PASCAL challenge [7].