Research map / Detection and segmentation / Open-set and few-shot detection
MUSES: The Multi-Sensor Semantic Perception Dataset for Driving under Uncertainty
Tim Broedermann, David Brüggemann, Christos Sakaridis, Kevin Ta, Odysseas Liagouris, Jason Corkill, Luc Van Gool
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 event, radar and lidar inputs to the camera in a simple cross-attention fusion network (ablation) | Mask2Former Swin-T · MUSES test, all conditions | PQ: 46.9 -> 53.6 | Improved |
| add extra sensors to the camera for night scenes (ablation) | Mask2Former Swin-T · MUSES test, night | PQ: 39.4 -> 49.7 | Improved |
| add one non-camera modality to the camera (each bimodal variant) (ablation) | Mask2Former Swin-T · MUSES test, all conditions | PQ: 46.9 -> 49.5 / 51.3 / 52.7 | Improved |
| use auxiliary modalities and normal-condition images in a second annotation stage (H2 vs H1) | MUSES annotation · MUSES | valid-label pixel coverage: increase of 10.3% to 78.5% valid pixels; about 1.5x more instances | Improved |
| evaluate on stage 2 labels instead of stage 1 labels | Mask2Former Swin-T · MUSES | PQ: 58.0 -> 46.9 | Hurt |
| train on MUSES instead of Cityscapes or ACDC | Mask2Former Swin-L · mean cross-dataset mIoU | mIoU: 69.4 / 71.4 -> 74.1 | Improved |
| extract confidence by marginalization instead of constant 100% confidence | Mask2Former Swin-T · MUSES uncertainty-aware panoptic | AUPQ (All): 46.9 -> 44.3 | Hurt |
| use higher-capacity segmentation backbone | MUSES semantic segmentation · MUSES test | mIoU: 70.7 (Swin-T) -> 77.1 (Swin-L) | Improved |
Closest papers by meaning
- ACDC: The Adverse Conditions Dataset With Correspondences for Robust Semantic Driving Scene Perception (TPAMI 2025 · similarity 0.90)
- STSBench: A Spatio-temporal Scenario Benchmark for Multi-modal Large Language Models in Autonomous Driving (NeurIPS 2025 · similarity 0.87)
- MCD: Diverse Large-Scale Multi-Campus Dataset for Robot Perception (CVPR 2024 · similarity 0.87)
- HiLoTs: High-Low Temporal Sensitive Representation Learning for Semi-Supervised LiDAR Segmentation in Autonomous Driving (CVPR 2025 · similarity 0.86)
- Panacea: Panoramic and Controllable Video Generation for Autonomous Driving (CVPR 2024 · similarity 0.86)
- The Road Ahead in Autonomous Driving: The KITScenes Multimodal Dataset (NeurIPS 2026 · 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.
