{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/106457"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/106457","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Differential treatment for stuff and things: A simple unsupervised domain adaptation method for semantic segmentation","abstract":"We consider the problem of unsupervised domain adaptation for semantic segmentation by easing the domain shift between the source domain (synthetic data) and the target domain (real data) in this work. State-of-the-art approaches prove that performing semantic-level alignment is helpful in tackling the domain shift issue. Based on the observation that stuff categories usually share similar appearances across images of different domains while things (i.e. object instances) have much larger differences, we propose to improve the semantic-level alignment with different strategies for stuff regions and for things: (1) for the stuff categories, we generate the feature representation for each class and conduct the alignment operation from the target domain to the source domain; (2) for the thing categories, we generate the feature representation for each individual instance and encourage the instance in the target domain to align with the most similar one in the source domain. In this way, the individual differences within thing categories will also be considered to alleviate over-alignment. In addition to our proposed method, we further reveal the reason why the current adversarial loss is often unstable in minimizing the distribution discrepancy and show that our method can help ease this issue by minimizing the most similar stuff and instance features between the source and the target domains. We conduct extensive experiments in two unsupervised domain adaptation tasks, GTA5 to Cityscapes and SYNTHIA to Cityscapes, and achieve the new state-of-the-art segmentation accuracy.","abstract_html":"We consider the problem of unsupervised domain adaptation for semantic segmentation by easing the domain shift between the source domain (synthetic data) and the target domain (real data) in this work. State-of-the-art approaches prove that performing semantic-level alignment is helpful in tackling the domain shift issue. Based on the observation that stuff categories usually share similar appearances across images of different domains while things (i.e. object instances) have much larger differences, we propose to improve the semantic-level alignment with different strategies for stuff regions and for things: (1) for the stuff categories, we generate the feature representation for each class and conduct the alignment operation from the target domain to the source domain; (2) for the thing categories, we generate the feature representation for each individual instance and encourage the instance in the target domain to align with the most similar one in the source domain. In this way, the individual differences within thing categories will also be considered to alleviate over-alignment. In addition to our proposed method, we further reveal the reason why the current adversarial loss is often unstable in minimizing the distribution discrepancy and show that our method can help ease this issue by minimizing the most similar stuff and instance features between the source and the target domains. We conduct extensive experiments in two unsupervised domain adaptation tasks, GTA5 to Cityscapes and SYNTHIA to Cityscapes, and achieve the new state-of-the-art segmentation accuracy.","abstract_has_math":false,"creators":["Wang, Zhonghao"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"M.S.","degree_level":"Thesis","degree_discipline":"Electrical & Computer Engr","degree_department":null,"school":null,"contributors":["Huang, Thomas S.","Shi, Honghui"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2020,"date_issued":"2020-03-02T22:38:46Z","date_published":"2020-03-02T22:38:46Z","updated_at":"2026-07-22T22:24:47Z","subjects":["domain adaptation","semantic segmentation"],"languages":["en"],"rights":["Copyright 2019 Zhonghao Wang"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/106457","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Huang, Thomas S.","Shi, Honghui"]},{"key":"dc:creator","label":"Author","values":["Wang, Zhonghao"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2020-03-02T22:38:46Z","2022-03-03T10:15:27Z","2019-11-22","2019-12"]},{"key":"dc:type","label":"Dc Type","values":["text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Electrical & Computer Engr"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Thesis"]},{"key":"thesis:degree_name","label":"Degree Name","values":["M.S."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois at Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["domain adaptation","semantic segmentation"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2019 Zhonghao Wang"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/106457"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["We consider the problem of unsupervised domain adaptation for semantic segmentation by easing the domain shift between the source domain (synthetic data) and the target domain (real data) in this work. State-of-the-art approaches prove that performing semantic-level alignment is helpful in tackling the domain shift issue. Based on the observation that stuff categories usually share similar appearances across images of different domains while things (i.e. object instances) have much larger differences, we propose to improve the semantic-level alignment with different strategies for stuff regions and for things: (1) for the stuff categories, we generate the feature representation for each class and conduct the alignment operation from the target domain to the source domain; (2) for the thing categories, we generate the feature representation for each individual instance and encourage the instance in the target domain to align with the most similar one in the source domain. In this way, the individual differences within thing categories will also be considered to alleviate over-alignment. In addition to our proposed method, we further reveal the reason why the current adversarial loss is often unstable in minimizing the distribution discrepancy and show that our method can help ease this issue by minimizing the most similar stuff and instance features between the source and the target domains. We conduct extensive experiments in two unsupervised domain adaptation tasks, GTA5 to Cityscapes and SYNTHIA to Cityscapes, and achieve the new state-of-the-art segmentation accuracy.","Submission published under a 24 month embargo labeled 'Closed Access', the embargo will last until 2021-12-01","The student, Zhonghao Wang, accepted the attached license on 2019-11-22 at 14:13.","The student, Zhonghao Wang, submitted this Thesis for approval on 2019-11-22 at 14:25.","This Thesis was approved for publication on 2019-11-22 at 15:53.","DSpace SAF Submission Ingestion Package generated from Vireo submission #14591 on 2020-02-28 at 17:36:44","Made available in DSpace on 2020-03-02T22:38:46Z (GMT). No. of bitstreams: 2 WANG-THESIS-2019.pdf: 1057263 bytes, checksum: c4adea20e1084d978a416296afffba84 (MD5) LICENSE.txt: 4210 bytes, checksum: c435e8b7ce1ce7c1d89116552470a39f (MD5) Previous issue date: 2019-11-22","Embargo set by: Seth Robbins for item 114001 Lift date: 2022-03-02T22:39:04Z Reason: Author requested closed access (OA after 2yrs) in Vireo ETD system","Limited Restriction Lifted for Item 114001 on 2022-03-03T10:15:27Z."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Differential treatment for stuff and things: A simple unsupervised domain adaptation method for semantic segmentation"]}]}],"canonical_facts":{"dc:contributor":["Huang, Thomas S.","Shi, Honghui"],"dc:creator":["Wang, Zhonghao"],"dc:date":["2020-03-02T22:38:46Z","2022-03-03T10:15:27Z","2019-11-22","2019-12"],"dc:description":["We consider the problem of unsupervised domain adaptation for semantic segmentation by easing the domain shift between the source domain (synthetic data) and the target domain (real data) in this work. State-of-the-art approaches prove that performing semantic-level alignment is helpful in tackling the domain shift issue. Based on the observation that stuff categories usually share similar appearances across images of different domains while things (i.e. object instances) have much larger differences, we propose to improve the semantic-level alignment with different strategies for stuff regions and for things: (1) for the stuff categories, we generate the feature representation for each class and conduct the alignment operation from the target domain to the source domain; (2) for the thing categories, we generate the feature representation for each individual instance and encourage the instance in the target domain to align with the most similar one in the source domain. In this way, the individual differences within thing categories will also be considered to alleviate over-alignment. In addition to our proposed method, we further reveal the reason why the current adversarial loss is often unstable in minimizing the distribution discrepancy and show that our method can help ease this issue by minimizing the most similar stuff and instance features between the source and the target domains. We conduct extensive experiments in two unsupervised domain adaptation tasks, GTA5 to Cityscapes and SYNTHIA to Cityscapes, and achieve the new state-of-the-art segmentation accuracy.","Submission published under a 24 month embargo labeled 'Closed Access', the embargo will last until 2021-12-01","The student, Zhonghao Wang, accepted the attached license on 2019-11-22 at 14:13.","The student, Zhonghao Wang, submitted this Thesis for approval on 2019-11-22 at 14:25.","This Thesis was approved for publication on 2019-11-22 at 15:53.","DSpace SAF Submission Ingestion Package generated from Vireo submission #14591 on 2020-02-28 at 17:36:44","Made available in DSpace on 2020-03-02T22:38:46Z (GMT). No. of bitstreams: 2 WANG-THESIS-2019.pdf: 1057263 bytes, checksum: c4adea20e1084d978a416296afffba84 (MD5) LICENSE.txt: 4210 bytes, checksum: c435e8b7ce1ce7c1d89116552470a39f (MD5) Previous issue date: 2019-11-22","Embargo set by: Seth Robbins for item 114001 Lift date: 2022-03-02T22:39:04Z Reason: Author requested closed access (OA after 2yrs) in Vireo ETD system","Limited Restriction Lifted for Item 114001 on 2022-03-03T10:15:27Z."],"dc:format":["application/pdf"],"dc:identifier":["http://hdl.handle.net/2142/106457"],"dc:language":["en"],"dc:rights":["Copyright 2019 Zhonghao Wang"],"dc:subject":["domain adaptation","semantic segmentation"],"dc:title":["Differential treatment for stuff and things: A simple unsupervised domain adaptation method for semantic segmentation"],"dc:type":["text"],"thesis:degree_discipline":["Electrical & Computer Engr"],"thesis:degree_level":["Thesis"],"thesis:degree_name":["M.S."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:24:47Z"}