WildScenes is a bi-modal benchmark dataset consisting of multiple large-scale, sequential traversals in natural environments, including semantic annotations in high-resolution 2D images and dense 3D LiDAR point clouds, and accurate 6-DoF pose information. The data is (1) trajectory-centric with accurate localization and globally aligned point clouds, (2) calibrated and synchronized to support bi-modal training and inference, and (3) containing different natural environments over 6 months to support research on domain adaptation. We introduce benchmarks on 2D and 3D semantic segmentation and evaluate a variety of recent deep-learning techniques to demonstrate the challenges in semantic segmentation in natural environments. We propose train-val-test splits for standard benchmarks as well as domain adaptation benchmarks and utilize an automated split generation technique to ensure the balance of class label distributions. The WildScenes benchmark webpage is https://csiro-robotics.github.io/WildScenes, and the data is publicly available at https://data.csiro.au/collection/csiro:61541 .
Variants: WildScenes
This dataset is used in 2 benchmarks:
| Task | Model | Paper | Date |
|---|---|---|---|
| 3D Semantic Segmentation | SphereFormer | Spherical Transformer for LiDAR-based 3D … | 2023-03-22 |
| 2D Semantic Segmentation | Mask2Former (ResNet-50) | Masked-attention Mask Transformer for Universal … | 2021-12-02 |
| 2D Semantic Segmentation | Mask2Former (Swin-L) | Masked-attention Mask Transformer for Universal … | 2021-12-02 |
| 2D Semantic Segmentation | Segformer (MiT-B5) | SegFormer: Simple and Efficient Design … | 2021-05-31 |
| 3D Semantic Segmentation | Cylinder3D | Cylinder3D: An Effective 3D Framework … | 2020-08-04 |
| 3D Semantic Segmentation | SPVCNN | Searching Efficient 3D Architectures with … | 2020-07-31 |
| 3D Semantic Segmentation | MinkUNet | 4D Spatio-Temporal ConvNets: Minkowski Convolutional … | 2019-04-18 |
| 2D Semantic Segmentation | UPerNet (ConvNeXt-L) | Unified Perceptual Parsing for Scene … | 2018-07-26 |
| 2D Semantic Segmentation | DeepLabv3 (ResNet-50) | Rethinking Atrous Convolution for Semantic … | 2017-06-17 |
Recent papers with results on this dataset: