ATLANTIS is a benchmark for semantic segmentation of waterbody images. This dataset covers a wide range of natural waterbodies such as sea, lake, river and man-made (artificial) water-related structures such as dam, reservoir, canal, and pier. ATLANTIS includes 5,195 pixel-wise annotated images split to 3,364 training, 535 validation, and 1,296 testing images. In addition to 35 waterbodies, this dataset covers 21 general labels such as person, car, road and building.
Variants: ATLANTIS
This dataset is used in 2 benchmarks:
| Task | Model | Paper | Date |
|---|---|---|---|
| Semantic Segmentation | Erfani et al. | ATLANTIS: A Benchmark for Semantic … | 2021-11-22 |
Recent papers with results on this dataset: