Semantic Co-segmentation in Videos

Discoveringandsegmentingobjectsinvideosisachallengingtaskduetolargevariationsofobjectsinappearances,deformedshapesandclutteredbackgrounds.Inthispaper,weproposetosegmentobjectsandunderstandtheirvisualsemanticsfromacollectionofvideosthatlinktoeachother,whichwerefertoassemanticco-segm