{"pk":28941,"title":"Robustness of Object Recognition under Extreme Occlusionin Humans and Computational Models","subtitle":null,"abstract":"Most objects in the visual world are partially occluded, buthumans can recognize them without difficulty. However, it re-mains unknown whether object recognition models like convo-lutional neural networks (CNNs) can handle real-world occlu-sion. It is also a question whether efforts to make these modelsrobust to constant mask occlusion are effective for real-worldocclusion. We test both humans and the above-mentionedcomputational models in a challenging task of object recogni-tion under extreme occlusion, where target objects are heavilyoccluded by irrelevant real objects in real backgrounds. Ourresults show that human vision is very robust to extreme oc-clusion while CNNs are not, even with modifications to han-dle constant mask occlusion. This implies that the ability tohandle constant mask occlusion does not entail robustness toreal-world occlusion. As a comparison, we propose anothercomputational model that utilizes object parts/subparts in acompositional manner to build robustness to occlusion. Thisperforms significantly better than CNN-based models on ourtask with error patterns similar to humans. These findings sug-gest that testing under extreme occlusion can better reveal therobustness of visual recognition, and that the principle of com-position can encourage such robustness.","language":"eng","license":{"name":"","short_name":"","text":null,"url":""},"keywords":[{"word":"visual recognition; occlusion; computationalmodel; neural network; psychophysics"}],"section":"Papers with Poster Presentations","is_remote":true,"remote_url":"https://escholarship.org/uc/item/0d5666gw","frozenauthors":[{"first_name":"Hongru","middle_name":"","last_name":"Zhu","name_suffix":"","institution":"The Johns Hopkins University, Baltimore","department":""},{"first_name":"Peng","middle_name":"","last_name":"Tang","name_suffix":"","institution":"The Johns Hopkins University, Baltimore","department":""},{"first_name":"Jeongho","middle_name":"","last_name":"Park","name_suffix":"","institution":"Harvard University","department":""},{"first_name":"Soojin","middle_name":"","last_name":"Park","name_suffix":"","institution":"Yonsei University","department":""},{"first_name":"Alan","middle_name":"","last_name":"Yuille","name_suffix":"","institution":"The Johns Hopkins University, Baltimore","department":""}],"date_submitted":null,"date_accepted":null,"date_published":"2019-01-01T18:00:00Z","render_galley":null,"galleys":[{"label":"PDF","type":"pdf","path":"https://journalpub.escholarship.org/cognitivesciencesociety/article/28941/galley/18812/download/"}]}