Research Article
Self-Recurrent Learning and Gap Sample Feature Synthesis-Based Object Detection Method
Table 1
Detection results on the COCO2017 dataset.
| Model | AP | [email protected] | [email protected] | APsmall | APmedium | APlarge |
| Libra R-CNN [16] | 43.0 | 64 | 47 | 25.3 | 45.6 | 54.6 | M2Det [38] | 41.0 | 59.7 | 45 | 22.1 | 46.5 | 53.8 | Faster R-CNN [21] | 40.6 | 58.9 | 44.5 | 22.0 | 42.8 | 52.6 | Cascade R-CNN [39] | 36.5 | 59 | 39.2 | 20.3 | 38.8 | 46.4 | RefineDet512 [40] | 33.0 | 54.5 | 35.5 | 16.3 | 36.3 | 44.3 | Ours | 43.7 | 60.5 | 49.1 | 28.0 | 52.0 | 59.4 |
|
|
We compared object detection results on five state-of-the-art methods to demonstrate the effectiveness of our method.
|