Object Detection on PASCAL VOC Novel Set 1 2007 (test)
80.8Base AP50Train base only
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Train base only# shots=3, Source=Our Impl.2020.03 | 80.8 | — | 9 | |
| Train base only# shots=102020.03 | 80.8 | — | 9 | |
| TFA w/ cos# shots=32020.03 | 79.1 | — | 44.7 | |
| TFA w/ cos# shots=102020.03 | 78.4 | — | 56 | |
| Meta R-CNN# shots=10, Source=Yan et al., 20192020.03 | 67.9 | — | 51.5 | |
| FRCN+ft-full# shots=3, Source=Our Impl.2020.03 | 66.1 | — | 29 | |
| FRCN+ft-full# shots=10, Source=Our Impl.2020.03 | 66 | — | 45.5 | |
| Meta R-CNN# shots=3, Source=Yan et al., 20192020.03 | 64.8 | — | 35 | |
| FRCN+ft-full# shots=3, Source=Yan et al., 20192020.03 | 63.6 | — | 32.8 | |
| FRCN+ft-full# shots=10, Source=Yan et al., 20192020.03 | 61.3 | — | 45.6 | |
| FewX+Oursbackbone=R101, shots=1, training=meta-learning2021.09 | — | 31.1 | — | |
| FsDetView+Ours bbackbone=R101, shots=10, training=meta-learning, multiple-run=true2021.09 | — | 59.2 | — | |
| LSTDbackbone=Darknet19, shots=1, training=meta-learning2021.09 | — | 8.2 | — | |
| NP-RepMetbackbone=R101DCN-FPN, shots=5, training=meta-learning2021.09 | — | 47.3 | — | |
| TFA w/ cos * bbackbone=R101FPN, shots=10, training=finetuning, multiple-run=true2021.09 | — | 52.8 | — |