Tiny Object Detection in Aerial Images
AI-TOD comes with 700,621 object instances for eight categories across 28,036 aerial images. Compared to existing object detection datasets in aerial images, the mean size of objects in AI-TOD is about 12.8 pixels, which is much smaller than others.
Variants: AI-TOD
This dataset is used in 2 benchmarks:
| Task | Model | Paper | Date |
|---|---|---|---|
| Object Detection | DNTR | A DeNoising FPN With Transformer … | 2024-06-09 |
| Object Detection | YOLOv5+Inner-IoU | Inner-IoU: More Effective Intersection over … | 2023-11-06 |
| Object Detection | DetectoRS w/ RFLA | RFLA: Gaussian Receptive Field based … | 2022-08-18 |
| Object Detection | DetectoRS + NWD (ResNet-50-FPN) | A Normalized Gaussian Wasserstein Distance … | 2021-10-26 |
| Object Detection | DetectoRS (ResNet-50-FPN) | DetectoRS: Detecting Objects with Recursive … | 2020-06-03 |
| Semantic Segmentation | Unet++(ResNet-50) | UNet++: A Nested U-Net Architecture … | 2018-07-18 |
| Semantic Segmentation | DeepLabV3+(ResNet-50) | Encoder-Decoder with Atrous Separable Convolution … | 2018-02-07 |
| Object Detection | Cascade R-CNN (ResNet-50-FPN) | Cascade R-CNN: Delving into High … | 2017-12-03 |
Recent papers with results on this dataset: