Abstract
Less than 35% of recyclable waste is being actually recycled in the US, whichleads to increased soil and sea pollution and is one of the major concerns ofenvironmental researchers as well as the common public. At the heart of theproblem are the inefficiencies of the waste sorting process (separating paper,plastic, metal, glass, etc.) due to the extremely complex and cluttered natureof the waste stream. Recyclable waste detection poses a unique computer visionchallenge as it requires detection of highly deformable and often translucentobjects in cluttered scenes without the kind of context information usuallypresent in human-centric datasets. This challenging computer vision taskcurrently lacks suitable datasets or methods in the available literature. Inthis paper, we take a step towards computer-aided waste detection and presentthe first in-the-wild industrial-grade waste detection and segmentationdataset, ZeroWaste. We believe that ZeroWaste will catalyze research in objectdetection and semantic segmentation in extreme clutter as well as applicationsin the recycling domain. Our project page can be found athttp://ai.bu.edu/zerowaste/.