Research map / Detection and segmentation / 3D detection and driving
Ithaca365: Dataset and Driving Perception Under Repeated and Challenging Weather Conditions
Carlos A. Diaz-Ruiz, Youya Xia, Yurong You, Jose Nino, Junan Chen, Josephine Monica, Xiangyu Chen, Katie Luo, Yan Wang, Marc Emond, Wei-Lun Chao, Bharath Hariharan and 2 more
Check your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add PAM, CAM, Mixture Pooling, and the residual sum operation to SFI | Amodal road segmentation · Ithaca365 road test | mIOU: 89.50 -> 92.19 | Improved |
| Train SFI on Ithaca365 rather than Cityscapes | SFI · KITTI road test | mIOU: 66.29 -> 72.31 | Improved |
| Train SFI on Ithaca365 rather than Cityscapes | SFI · Cityscapes road test | mIOU: 72.75 -> 74.68 | Improved |
| Train the road segmentation model on Ithaca365 rather than Cityscapes | SFI · Ithaca365 road test | mIOU: 50.16 -> 89.50 | Improved |
| Train and test on the same weather condition rather than a different weather condition | Mask R-CNN amodal instance segmentation · Ithaca365 weather-specific tests | mAP@[.5:.95]: Table 4 diagonal entries are the highest for each test-weather column | Improved |
| Train PSMNet on the test weather rather than another weather | PSMNet stereo matching · Ithaca365 rain, snow, and night | Three-pixel-error (%): Table 6 diagonal errors are usually much smaller than cross-domain errors | Improved |
| Use the proposed amodal road model rather than SFI | Amodal road segmentation · KITTI road test | mIOU: 72.31 -> 72.25 | No clear effect |
| Use Mixture Pooling instead of Max Pooling as Inpainting (ablation) | Amodal road segmentation · Qualitative inpainting results | inpainting boundary smoothness | Improved |
| Add a residual sum link from the PAM output before the final up-sampling layer (ablation) | Amodal road segmentation · Qualitative inpainting results | inpainting quality | Improved |
| Add the residual sum link from PAM (ablation) | Amodal road segmentation · Images where foreground objects occupy a relatively large space | inpainting quality | Improved |
Closest papers by meaning
- One Million Scenes for Autonomous Driving: ONCE Dataset (NeurIPS 2021 · similarity 0.90)
- Scalability in Perception for Autonomous Driving: Waymo Open Dataset (CVPR 2020 · similarity 0.89)
- MSU-4S - The Michigan State University Four Seasons Dataset (CVPR 2024 · similarity 0.88)
- The Road Ahead in Autonomous Driving: The KITScenes Multimodal Dataset (NeurIPS 2026 · similarity 0.87)
- V2X-Radar: A Multi-modal Dataset with 4D Radar for Cooperative Perception (NeurIPS 2025 · similarity 0.87)
- Spotting the Unexpected (STU): A 3D LiDAR Dataset for Anomaly Segmentation in Autonomous Driving (CVPR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
