Research map / Detection and segmentation / CNN features
MARSS: Radar Semantic Segmentation via Modular Attention and State Space Models
Fengyu Chen, Tiao Tan, Teng Li, Yuantian Quan, Qingmin Liao
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 RADE channel and spatial attention encoder (ablation) | MARSS baseline · CARRADA RD view | mIoU: 61.16% -> 62.35% | Improved |
| add RADE encoder module (ablation) | MARSS baseline · CARRADA RA view | mIoU: 41.04% -> 41.57% | No clear effect |
| add RFAF multi-scale and region attention fusion (ablation) | MARSS baseline · CARRADA RA view | mIoU: 41.04% -> 44.70% | Improved |
| add RADM Mamba and axial attention decoder (ablation) | MARSS baseline · CARRADA RA view | mIoU: 41.04% -> 45.33% | Improved |
| add RADM decoder (ablation) | MARSS baseline · CARRADA RD view | mIoU: 61.16% -> 62.52% | Improved |
| combine all three modules RADE, RFAF, RADM (ablation) | MARSS · CARRADA RA view | mIoU: 45.33% -> 46.97% | Improved |
| full MARSS versus no modules (ablation) | MARSS · CARRADA RD view | mIoU: 61.16% -> 63.26% | Improved |
| MARSS versus AdaPKC-xi-Net FiT | MARSS · CARRADA RA view | mIoU: 44.30% -> 46.97% | Improved |
| MARSS versus AdaPKC-xi-Net on RD view | MARSS · CARRADA-RAC RD view | mIoU: 61.6% -> 61.4% | No clear effect |
| MARSS versus AdaPKC-xi-Net on global view | MARSS · CARRADA-RAC Global view | mIoU: 54.8% -> 55.2% | Improved |
Closest papers by meaning
- TARSS-Net: Temporal-Aware Radar Semantic Segmentation Network (NeurIPS 2024 · similarity 0.87)
- AdaPKC: PeakConv with Adaptive Peak Receptive Field for Radar Semantic Segmentation (NeurIPS 2024 · similarity 0.86)
- RETR: Multi-View Radar Detection Transformer for Indoor Perception (NeurIPS 2024 · similarity 0.84)
- RadarMOSEVE: A Spatial-Temporal Transformer Network for Radar-Only Moving Object Segmentation and Ego-Velocity Estimation (AAAI 2024 · similarity 0.84)
- Selection, Not Fusion: Radar-Modulated State Space Models for Radar-Camera Depth Estimation (NeurIPS 2026 · similarity 0.83)
- RADAR: Text-Guided Medical Image Segmentation via Residual Aggregation and Dense Alignment Representations (NeurIPS 2026 · 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.
