Research map / Detection and segmentation / 3D detection and driving
SpaceNet MVOI: A Multi-View Overhead Imagery Dataset
Nicholas Weir, David Lindenbaum, Alexei Bastidas, Adam Van Etten, Sean McPherson, Jacob Shermeyer, Varun Kumar, Hanlin Tang
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 |
|---|---|---|---|
| train on images down-sampled to a common 1.67 m resolution instead of original resolution (ablation) | TernausNet · SpaceNet MVOI building segmentation | F1: 0.43 -> 0.41 | No clear effect |
| train only on NADIR views instead of training on OFF views, tested on OFF | TernausNet · SpaceNet MVOI OFF test | F1: 0.44 -> 0.32 | Hurt |
| test on very off-nadir instead of nadir views (trained on all angles) | TernausNet · SpaceNet MVOI building segmentation | F1: 0.62 -> 0.22 | Hurt |
| use overhead-specific model TernausNet instead of general U-Net | segmentation · SpaceNet MVOI average F1 | F1: 0.24 -> 0.43 | Improved |
| use overhead-specific YOLT instead of Mask R-CNN detection | object detection · SpaceNet MVOI average F1 | F1: 0.25 -> 0.36 | Improved |
| use Mask R-CNN instead of TernausNet in terms of generalization to unseen angles | segmentation · SpaceNet MVOI | generalization score G: 0.42 -> 0.78 | Improved |
| train on NADIR only, tested on VOFF instead of training on all angles | TernausNet · SpaceNet MVOI VOFF test | F1: 0.22 -> 0.04 | Hurt |
| test on unseen Las Vegas geography instead of MVOI test set | TernausNet trained on MVOI ALL · SpaceNet cross-dataset | F1: 0.68 -> 0.01 | Hurt |
Closest papers by meaning
- The Multi-Temporal Urban Development SpaceNet Dataset (CVPR 2021 · similarity 0.86)
- DynamicVL: Benchmarking Multimodal Large Language Models for Dynamic City Understanding (NeurIPS 2025 · similarity 0.85)
- TEOChat: A Large Vision-Language Assistant for Temporal Earth Observation Data (ICLR 2025 · similarity 0.85)
- Unleashing Multispectral Video's Potential in Semantic Segmentation: A Semi-supervised Viewpoint and New UAV-View Benchmark (NeurIPS 2024 · similarity 0.84)
- Benchmarking Object Detectors under Real-World Distribution Shifts in Satellite Imagery (CVPR 2025 · similarity 0.84)
- Recognize Any Regions (NeurIPS 2024 · similarity 0.84)
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.
