Frustratingly Simple Few-Shot Object Detection

Xin Wang, Thomas Huang, Joseph Gonzalez, Trevor Darrell, Fisher Yu
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:9919-9928, 2020.

Abstract

Detecting rare objects from a few examples is an emerging problem. Prior works show meta-learning is a promising approach. But, fine-tuning techniques have drawn scant attention. We find that fine-tuning only the last layer of existing detectors on rare classes is crucial to the few-shot object detection task. Such a simple approach outperforms the meta-learning methods by roughly 2 20 points on current benchmarks and sometimes even doubles the accuracy of the prior methods. However, the high variance in the few samples often leads to the unreliability of existing benchmarks. We revise the evaluation protocols by sampling multiple groups of training examples to obtain stable comparisons and build new benchmarks based on three datasets: PASCAL VOC, COCO and LVIS. Again, our fine-tuning approach establishes a new state of the art on the revised benchmarks. The code as well as the pretrained models are available at https://github.com/ucbdrive/few-shot-object-detection.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-wang20j, title = {Frustratingly Simple Few-Shot Object Detection}, author = {Wang, Xin and Huang, Thomas and Gonzalez, Joseph and Darrell, Trevor and Yu, Fisher}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {9919--9928}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/wang20j/wang20j.pdf}, url = {https://proceedings.mlr.press/v119/wang20j.html}, abstract = {Detecting rare objects from a few examples is an emerging problem. Prior works show meta-learning is a promising approach. But, fine-tuning techniques have drawn scant attention. We find that fine-tuning only the last layer of existing detectors on rare classes is crucial to the few-shot object detection task. Such a simple approach outperforms the meta-learning methods by roughly 2 20 points on current benchmarks and sometimes even doubles the accuracy of the prior methods. However, the high variance in the few samples often leads to the unreliability of existing benchmarks. We revise the evaluation protocols by sampling multiple groups of training examples to obtain stable comparisons and build new benchmarks based on three datasets: PASCAL VOC, COCO and LVIS. Again, our fine-tuning approach establishes a new state of the art on the revised benchmarks. The code as well as the pretrained models are available at https://github.com/ucbdrive/few-shot-object-detection.} }
Endnote
%0 Conference Paper %T Frustratingly Simple Few-Shot Object Detection %A Xin Wang %A Thomas Huang %A Joseph Gonzalez %A Trevor Darrell %A Fisher Yu %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-wang20j %I PMLR %P 9919--9928 %U https://proceedings.mlr.press/v119/wang20j.html %V 119 %X Detecting rare objects from a few examples is an emerging problem. Prior works show meta-learning is a promising approach. But, fine-tuning techniques have drawn scant attention. We find that fine-tuning only the last layer of existing detectors on rare classes is crucial to the few-shot object detection task. Such a simple approach outperforms the meta-learning methods by roughly 2 20 points on current benchmarks and sometimes even doubles the accuracy of the prior methods. However, the high variance in the few samples often leads to the unreliability of existing benchmarks. We revise the evaluation protocols by sampling multiple groups of training examples to obtain stable comparisons and build new benchmarks based on three datasets: PASCAL VOC, COCO and LVIS. Again, our fine-tuning approach establishes a new state of the art on the revised benchmarks. The code as well as the pretrained models are available at https://github.com/ucbdrive/few-shot-object-detection.
APA
Wang, X., Huang, T., Gonzalez, J., Darrell, T. & Yu, F.. (2020). Frustratingly Simple Few-Shot Object Detection. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:9919-9928 Available from https://proceedings.mlr.press/v119/wang20j.html.

Related Material