Research map / Detection and segmentation / Open-set and few-shot detection
The Mapillary Vistas Dataset for Semantic Understanding of Street Scenes
Gerhard Neuhold, Tobias Ollmann, Samuel Rota Bulo, Peter Kontschieder
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 loss max-pooling | Wider Network · Mapillary Vistas semantic segmentation (val) | mean IoU: 41.12 -> 43.78 | Improved |
| add balanced class sampling | Wider Network + loss max-pooling · Mapillary Vistas semantic segmentation (val) | mean IoU: 43.78 -> 47.72 | Improved |
| add balanced class sampling | Wider Network + loss max-pooling · Mapillary Vistas semantic segmentation (test) | mean IoU: 42.98 -> 44.84 | Improved |
| add HDC and auxiliary loss | PSPNet · Mapillary Vistas semantic segmentation (val) | mean IoU: 49.76 -> 50.28 | Improved |
| add class reweighting | PSPNet · Mapillary Vistas semantic segmentation (val) | mean IoU: 50.28 -> 51.50 | Improved |
| add Cityscapes pre-training | PSPNet · Mapillary Vistas semantic segmentation (val) | mean IoU: 51.50 -> 51.59 | No clear effect |
| add multi-scale testing with 6 scales | PSPNet · Mapillary Vistas semantic segmentation (val) | mean IoU: 51.59 -> 53.51 | Improved |
| ensemble 4 models | PSPNet · Mapillary Vistas semantic segmentation (val) | mean IoU: 53.51 -> 53.85 | Improved |
| ensemble of models instead of single model | UCenter Mask R-CNN extension · Mapillary Vistas instance segmentation (val) | mask AP: 22.8 -> 23.7 | Improved |
| downsample label images by a larger factor (2 to 128) | label upper bound · Mapillary Vistas category-level labels | IoU upper bound: 98.00 -> 50.67 | Hurt |
| use coarser root-level labels instead of category-level at downsampling factor 128 | label upper bound · Mapillary Vistas | IoU upper bound: 50.67 -> 63.85 | Improved |
Closest papers by meaning
- The Cityscapes Dataset for Semantic Urban Scene Understanding (CVPR 2016 · similarity 0.88)
- The Mapillary Traffic Sign Dataset for Detection and Classification on a Global Scale (ECCV 2020 · similarity 0.87)
- Mapillary Street-Level Sequences: A Dataset for Lifelong Place Recognition (CVPR 2020 · similarity 0.87)
- Weakly Supervised Semantic Segmentation for Driving Scenes (AAAI 2024 · similarity 0.84)
- USCILab3D: A Large-scale, Long-term, Semantically Annotated Outdoor Dataset (NeurIPS 2024 · similarity 0.84)
- SUM Parts: Benchmarking Part-Level Semantic Segmentation of Urban Meshes (CVPR 2025 · 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.
