TY - GEN
T1 - DETReg
T2 - 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2022
AU - Bar, Amir
AU - Wang, Xin
AU - Kantorov, Vadim
AU - Reed, Colorado J.
AU - Herzig, Roei
AU - Chechik, Gal
AU - Rohrbach, Anna
AU - Darrell, Trevor
AU - Globerson, Amir
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - Recent self-supervised pretraining methods for object detection largely focus on pretraining the backbone of the object detector, neglecting key parts of detection architecture. Instead, we introduce DETReg, a new self-supervised method that pretrains the entire object detection network, including the object localization and embedding components. During pretraining, DETReg predicts object localizations to match the localizations from an unsupervised region proposal generator and simultaneously aligns the corresponding feature embeddings with embeddings from a self-supervised image encoder. We implement DETReg using the DETR family of detectors and show that it improves over competitive baselines when finetuned on COCO, PASCAL VOC, and Airbus Ship benchmarks. In low-data regimes, including semi-supervised and few-shot learning settings, DETReg establishes many state-of-the-art results, e.g., on COCO we see a +6.0 AP improvement for 10-shot detection and over 2 AP improvements when training with only 1 % of the labels.11Code: https://www.amirbar.net/detreg/.
AB - Recent self-supervised pretraining methods for object detection largely focus on pretraining the backbone of the object detector, neglecting key parts of detection architecture. Instead, we introduce DETReg, a new self-supervised method that pretrains the entire object detection network, including the object localization and embedding components. During pretraining, DETReg predicts object localizations to match the localizations from an unsupervised region proposal generator and simultaneously aligns the corresponding feature embeddings with embeddings from a self-supervised image encoder. We implement DETReg using the DETR family of detectors and show that it improves over competitive baselines when finetuned on COCO, PASCAL VOC, and Airbus Ship benchmarks. In low-data regimes, including semi-supervised and few-shot learning settings, DETReg establishes many state-of-the-art results, e.g., on COCO we see a +6.0 AP improvement for 10-shot detection and over 2 AP improvements when training with only 1 % of the labels.11Code: https://www.amirbar.net/detreg/.
KW - Representation learning
KW - Self-& semi-& meta- Recognition: detection
KW - categorization
KW - retrieval
UR - http://www.scopus.com/inward/record.url?scp=85132059207&partnerID=8YFLogxK
U2 - 10.1109/CVPR52688.2022.01420
DO - 10.1109/CVPR52688.2022.01420
M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???
AN - SCOPUS:85132059207
T3 - Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
SP - 14585
EP - 14595
BT - Proceedings - 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2022
PB - IEEE Computer Society
Y2 - 19 June 2022 through 24 June 2022
ER -