LEDetection: A Simple Framework for Semi-Supervised Few-Shot Object Detection

Phi Vu Tran
Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, PMLR 238:640-648, 2024.

Abstract

Few-shot object detection (FSOD) is a challenging problem aimed at detecting novel concepts from few exemplars. Existing approaches to FSOD all assume abundant base labels to adapt to novel objects. This paper studies the new task of semi-supervised FSOD by considering a realistic scenario in which both base and novel labels are simultaneously scarce. We explore the utility of unlabeled data within our proposed label-efficient detection framework and discover its remarkable ability to boost semi-supervised FSOD by way of region proposals. Motivated by this finding, we introduce SoftER Teacher, a robust detector combining pseudo-labeling with consistency learning on region proposals, to harness unlabeled data for improved FSOD without relying on abundant labels. Rigorous experiments show that SoftER Teacher surpasses the novel performance of a strong supervised detector using only 10% of required base labels, without catastrophic forgetting observed in prior approaches. Our work also sheds light on a potential relationship between semi-supervised and few-shot detection suggesting that a stronger semi-supervised detector leads to a more effective few-shot detector.

Cite this Paper


BibTeX
@InProceedings{pmlr-v238-vu-tran24a, title = {{LEDetection}: A Simple Framework for Semi-Supervised Few-Shot Object Detection}, author = {Vu Tran, Phi}, booktitle = {Proceedings of The 27th International Conference on Artificial Intelligence and Statistics}, pages = {640--648}, year = {2024}, editor = {Dasgupta, Sanjoy and Mandt, Stephan and Li, Yingzhen}, volume = {238}, series = {Proceedings of Machine Learning Research}, month = {02--04 May}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v238/vu-tran24a/vu-tran24a.pdf}, url = {https://proceedings.mlr.press/v238/vu-tran24a.html}, abstract = {Few-shot object detection (FSOD) is a challenging problem aimed at detecting novel concepts from few exemplars. Existing approaches to FSOD all assume abundant base labels to adapt to novel objects. This paper studies the new task of semi-supervised FSOD by considering a realistic scenario in which both base and novel labels are simultaneously scarce. We explore the utility of unlabeled data within our proposed label-efficient detection framework and discover its remarkable ability to boost semi-supervised FSOD by way of region proposals. Motivated by this finding, we introduce SoftER Teacher, a robust detector combining pseudo-labeling with consistency learning on region proposals, to harness unlabeled data for improved FSOD without relying on abundant labels. Rigorous experiments show that SoftER Teacher surpasses the novel performance of a strong supervised detector using only 10% of required base labels, without catastrophic forgetting observed in prior approaches. Our work also sheds light on a potential relationship between semi-supervised and few-shot detection suggesting that a stronger semi-supervised detector leads to a more effective few-shot detector.} }
Endnote
%0 Conference Paper %T LEDetection: A Simple Framework for Semi-Supervised Few-Shot Object Detection %A Phi Vu Tran %B Proceedings of The 27th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2024 %E Sanjoy Dasgupta %E Stephan Mandt %E Yingzhen Li %F pmlr-v238-vu-tran24a %I PMLR %P 640--648 %U https://proceedings.mlr.press/v238/vu-tran24a.html %V 238 %X Few-shot object detection (FSOD) is a challenging problem aimed at detecting novel concepts from few exemplars. Existing approaches to FSOD all assume abundant base labels to adapt to novel objects. This paper studies the new task of semi-supervised FSOD by considering a realistic scenario in which both base and novel labels are simultaneously scarce. We explore the utility of unlabeled data within our proposed label-efficient detection framework and discover its remarkable ability to boost semi-supervised FSOD by way of region proposals. Motivated by this finding, we introduce SoftER Teacher, a robust detector combining pseudo-labeling with consistency learning on region proposals, to harness unlabeled data for improved FSOD without relying on abundant labels. Rigorous experiments show that SoftER Teacher surpasses the novel performance of a strong supervised detector using only 10% of required base labels, without catastrophic forgetting observed in prior approaches. Our work also sheds light on a potential relationship between semi-supervised and few-shot detection suggesting that a stronger semi-supervised detector leads to a more effective few-shot detector.
APA
Vu Tran, P.. (2024). LEDetection: A Simple Framework for Semi-Supervised Few-Shot Object Detection. Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 238:640-648 Available from https://proceedings.mlr.press/v238/vu-tran24a.html.

Related Material