gaobb/Few-Shot-Object-Detection-Papers

Three missing FSOD papers

GuangxingHan opened this issue · 3 comments

gaobb commented

Thanks very much for pointing out the 3 missing papers.
I am very grateful if you can provide the results of the corresponding methods on the MS-COCO and PASCAL-VOC datasets in the format of our leaderboard.

Thanks.

The MSCOCO results are:
|CoCo-RCNN| ECCV |2022|R-101|Sparse-RCNN| Fine-tuning |FSOD|5.2/-/16.4/19.2|PyTorch|
|Meta-Faster-RCNN| AAAI |2022|R-101|Faster R-CNN| meta-learning|FSOD|5.1/10.8/12.7/16.6|PyTorch|
|QA-FewDet| ICCV |2021|R-101|Faster R-CNN| meta-learning|FSOD|4.9/9.7/11.6/16.5|PyTorch|

The VOC results are:
|CoCo-RCNN| ECCV |2022|R-101|Sparse-RCNN| Fine-tuning |FSOD|33.5 44.2 50.2 57.5 63.3 | 25.3 31.0 39.6 43.8 50.1 | 24.8 36.9 42.8 50.8 57.7|PyTorch|
|Meta-Faster-RCNN| AAAI |2022|R-101|Faster R-CNN| meta-learning|FSOD|43.0 54.5 60.6 66.1 65.4 |27.7 35.5 46.1 47.8 51.4 |40.6 46.4 53.4 59.9 58.6|PyTorch|
|QA-FewDet| ICCV |2021|R-101|Faster R-CNN| meta-learning|FSOD|42.4 51.9 55.7 62.6 63.4| 25.9 37.8 46.6 48.9 51.1| 35.2 42.9 47.8 54.8 53.5|PyTorch|

By the way, our FCT (CVPR 2022) has the open-sourced project at https://github.com/GuangxingHan/FCT.

gaobb commented

Thanks very much.
The provided results and the corresponding papers have been updated. In addition, the FCT code is also updated.