{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/109535"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/109535","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Improving few-shot object detection by saving and hallucinating examples","abstract":"Learning to detect an object in an image from very few training examples - few-shot object detection - is challenging, because the classifier that sees proposal boxes has very little training data. A particularly challenging training regime occurs when there are one or two training examples. In this case, if the region proposal network (RPN) misses even one high intersection-over-union (IOU) training box, the classifier's model of how object appearance varies can be severely impacted. We use multiple distinct yet cooperating RPN's. Our RPN's are trained to be different, but not too different; doing so yields significant performance improvements over state of the art for COCO and PASCAL VOC in the very few-shot setting. This effect appears to be independent of the choice of classifier or dataset. However, under the very low-shot regime, even if all high IOU boxes are used to train the classifier, the variations are still insufficient to train the classifier in novel classes. We propose to build an even better model of variation in novel classes by transferring the shared within-class variation from base classes. We introduce a hallucinator network and insert it into a modern object detector model, which learns to generate additional training examples in the Region of Interest (ROI's) feature space. This approach yields further performance improvements on two state-of-the-art few-shot detectors with different proposal generation processes.","abstract_html":"Learning to detect an object in an image from very few training examples - few-shot object detection - is challenging, because the classifier that sees proposal boxes has very little training data. A particularly challenging training regime occurs when there are one or two training examples. In this case, if the region proposal network (RPN) misses even one high intersection-over-union (IOU) training box, the classifier's model of how object appearance varies can be severely impacted. We use multiple distinct yet cooperating RPN's. Our RPN's are trained to be different, but not too different; doing so yields significant performance improvements over state of the art for COCO and PASCAL VOC in the very few-shot setting. This effect appears to be independent of the choice of classifier or dataset. However, under the very low-shot regime, even if all high IOU boxes are used to train the classifier, the variations are still insufficient to train the classifier in novel classes. We propose to build an even better model of variation in novel classes by transferring the shared within-class variation from base classes. We introduce a hallucinator network and insert it into a modern object detector model, which learns to generate additional training examples in the Region of Interest (ROI's) feature space. This approach yields further performance improvements on two state-of-the-art few-shot detectors with different proposal generation processes.","abstract_has_math":false,"creators":["Zhang, Weilin"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"M.S.","degree_level":"Thesis","degree_discipline":"Computer Science","degree_department":null,"school":null,"contributors":["Forsyth, David Alexander"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2021,"date_issued":"2021-03-05T21:42:52Z","date_published":"2021-03-05T21:42:52Z","updated_at":"2026-07-22T22:24:50Z","subjects":["Object Detection","Few-Shot Learning"],"languages":["en"],"rights":["Copyright 2020 Weilin Zhang"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/109535","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Forsyth, David Alexander"]},{"key":"dc:creator","label":"Author","values":["Zhang, Weilin"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2021-03-05T21:42:52Z","2023-03-05T21:43:00Z","2020-12-09","2020-12"]},{"key":"dc:type","label":"Dc Type","values":["text","Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science"]},{"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":["Object Detection","Few-Shot Learning"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2020 Weilin Zhang"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/109535"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Learning to detect an object in an image from very few training examples - few-shot object detection - is challenging, because the classifier that sees proposal boxes has very little training data. A particularly challenging training regime occurs when there are one or two training examples. In this case, if the region proposal network (RPN) misses even one high intersection-over-union (IOU) training box, the classifier's model of how object appearance varies can be severely impacted. We use multiple distinct yet cooperating RPN's. Our RPN's are trained to be different, but not too different; doing so yields significant performance improvements over state of the art for COCO and PASCAL VOC in the very few-shot setting. This effect appears to be independent of the choice of classifier or dataset. However, under the very low-shot regime, even if all high IOU boxes are used to train the classifier, the variations are still insufficient to train the classifier in novel classes. We propose to build an even better model of variation in novel classes by transferring the shared within-class variation from base classes. We introduce a hallucinator network and insert it into a modern object detector model, which learns to generate additional training examples in the Region of Interest (ROI's) feature space. This approach yields further performance improvements on two state-of-the-art few-shot detectors with different proposal generation processes.","Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2022-12-01","The student, Weilin Zhang, accepted the attached license on 2020-12-08 at 16:29.","The student, Weilin Zhang, submitted this Thesis for approval on 2020-12-08 at 16:37.","This Thesis was approved for publication on 2020-12-09 at 15:03.","DSpace SAF Submission Ingestion Package generated from Vireo submission #16094 on 2021-03-04 at 16:20:45","Made available in DSpace on 2021-03-05T21:42:52Z (GMT). No. of bitstreams: 2 ZHANG-THESIS-2020.pdf: 23061019 bytes, checksum: 1aa422e775a95bd91020fc090a9fdcdc (MD5) LICENSE.txt: 4209 bytes, checksum: ae8e6496cf2d269d7ab7c3ef04755004 (MD5) Previous issue date: 2020-12-09","Embargo set by: Seth Robbins for item 117240 Lift date: 2023-03-05T21:43:00Z Reason: Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","U of I Only"]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Improving few-shot object detection by saving and hallucinating examples"]}]}],"canonical_facts":{"dc:contributor":["Forsyth, David Alexander"],"dc:creator":["Zhang, Weilin"],"dc:date":["2021-03-05T21:42:52Z","2023-03-05T21:43:00Z","2020-12-09","2020-12"],"dc:description":["Learning to detect an object in an image from very few training examples - few-shot object detection - is challenging, because the classifier that sees proposal boxes has very little training data. A particularly challenging training regime occurs when there are one or two training examples. In this case, if the region proposal network (RPN) misses even one high intersection-over-union (IOU) training box, the classifier's model of how object appearance varies can be severely impacted. We use multiple distinct yet cooperating RPN's. Our RPN's are trained to be different, but not too different; doing so yields significant performance improvements over state of the art for COCO and PASCAL VOC in the very few-shot setting. This effect appears to be independent of the choice of classifier or dataset. However, under the very low-shot regime, even if all high IOU boxes are used to train the classifier, the variations are still insufficient to train the classifier in novel classes. We propose to build an even better model of variation in novel classes by transferring the shared within-class variation from base classes. We introduce a hallucinator network and insert it into a modern object detector model, which learns to generate additional training examples in the Region of Interest (ROI's) feature space. This approach yields further performance improvements on two state-of-the-art few-shot detectors with different proposal generation processes.","Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2022-12-01","The student, Weilin Zhang, accepted the attached license on 2020-12-08 at 16:29.","The student, Weilin Zhang, submitted this Thesis for approval on 2020-12-08 at 16:37.","This Thesis was approved for publication on 2020-12-09 at 15:03.","DSpace SAF Submission Ingestion Package generated from Vireo submission #16094 on 2021-03-04 at 16:20:45","Made available in DSpace on 2021-03-05T21:42:52Z (GMT). No. of bitstreams: 2 ZHANG-THESIS-2020.pdf: 23061019 bytes, checksum: 1aa422e775a95bd91020fc090a9fdcdc (MD5) LICENSE.txt: 4209 bytes, checksum: ae8e6496cf2d269d7ab7c3ef04755004 (MD5) Previous issue date: 2020-12-09","Embargo set by: Seth Robbins for item 117240 Lift date: 2023-03-05T21:43:00Z Reason: Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","U of I Only"],"dc:format":["application/pdf"],"dc:identifier":["http://hdl.handle.net/2142/109535"],"dc:language":["en"],"dc:rights":["Copyright 2020 Weilin Zhang"],"dc:subject":["Object Detection","Few-Shot Learning"],"dc:title":["Improving few-shot object detection by saving and hallucinating examples"],"dc:type":["text","Thesis"],"thesis:degree_discipline":["Computer Science"],"thesis:degree_level":["Thesis"],"thesis:degree_name":["M.S."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:24:50Z"}