Hierarchical waste detection with weakly supervised segmentation in images from recycling plants

Dmitry A. Yudin, Nikita Zakharenko, Artem Smetanin, Roman Filonov, Margarita Kichik, Vladislav Kuznetsov, Dmitry Larichev, Evgeny Gudov, Semen A. Budennyy, Aleksandr Panov

Abstract: Highlights•We developed a large diverse waste dataset called WaRP. It contains more than 10,000 manually labeled images with 28 object types that can be found on the conveyor belt of recycling plants.•We proposed a novel architecture of a hierarchical neural network called H-YC for waste detection with weakly supervised segmentation.•Experiments on real data demonstrated its state-of-the-art quality.