Research map / Detection and segmentation / Open-set and few-shot detection
TorontoCity: Seeing the World With a Million Eyes
Shenlong Wang, Min Bai, Gellert Mattyus, Hang Chu, Wenjie Luo, Bin Yang, Justin Liang, Joel Cheverie, Sanja Fidler, Raquel Urtasun
Read the paperCheck 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 morphological opening before instance extraction (ablation) | FCN building instance seg · TorontoCity aerial | AP: 8.04% -> 16.45% | Improved |
| add morphological opening before instance extraction (ablation) | ResNet building instance seg · TorontoCity aerial | AP: 10.47% -> 22.92% | Improved |
| add morphological closing for road centerline (ablation) | ResNet · Road centerline | F1 at 2m: 0.613 -> 0.644 | Improved |
| add morphological closing for road centerline (ablation) | ResNet · Road centerline | F1 at 0.5m: 0.162 -> 0.162 | No clear effect |
| add morphological closing for road curb (ablation) | ResNet · Road curb | F1 at 0.5m: 0.575 -> 0.568 | Hurt |
| add morphological closing for road curb (ablation) | FCN · Road curb | F1 at 2m: 0.778 -> 0.781 | No clear effect |
| use ResNet instead of FCN for aerial segmentation | segmentation · Aerial mean IoU | IoU: 77.64% -> 78.46% | Improved |
| use ResNet instead of FCN for building contours | contour extraction · Building contour | WeightedCov: 0.456 -> 0.401 | Hurt |
| use ImageNet-pretrained ResNet-152 vs from-scratch ResNet-32 | urban zoning classifier · Ground zoning | top-1 accuracy: 75.65% -> 79.33% | Improved |
| estimate building height from aerial images | CNNs · Building height | log-RMSE | No clear effect |
Closest papers by meaning
- CityRefer: Geography-aware 3D Visual Grounding Dataset on City-scale Point Cloud Data (NeurIPS 2023 · similarity 0.85)
- The Cityscapes Dataset for Semantic Urban Scene Understanding (CVPR 2016 · similarity 0.85)
- Deep Learning the City: Quantifying Urban Perception at a Global Scale (ECCV 2016 · similarity 0.84)
- SensatUrban: Learning Semantics from Urban-Scale Photogrammetric Point Clouds (IJCV 2022 · similarity 0.84)
- GeoFidelity-Bench: Evaluating Block-Conditioned Geographic Fidelity of Street-View Generation (NeurIPS 2026 · similarity 0.84)
- UrbanDataLayer: A Unified Data Pipeline for Urban Science (NeurIPS 2024 · similarity 0.83)
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.
