Research map / Detection and segmentation / CNN features
Localized Semantic Feature Mixers for Efficient Pedestrian Detection in Autonomous Driving
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Replace the baseline detection head with DFDN (ablation) | HRNet-based detector · City Persons | MR−2: 9.5 -> 9.9 | Hurt |
| Add Fast Suppression Head (ablation) | HRNet-based detector with DFDN · City Persons | MR−2: 9.9 -> 9.5 | Helped |
| Add hard mixup augmentation (ablation) | HRNet-based detector with DFDN and FSH · City Persons | MR−2: 9.5 -> 9.2 | Helped |
| Add mean teacher (ablation) | HRNet-based detector with DFDN, FSH, and hard mixup · City Persons | MR−2: 9.2 -> 8.8 | Helped |
| Add SP3 (ablation) | HRNet-based detector with DFDN, FSH, hard mixup, and mean teacher · City Persons | MR−2: 8.8 -> 8.5 | Helped |
| Add hard mixup augmentation (ablation) | ConvMLP-Pin detector with DFDN and FSH · City Persons | MR−2: 10.5 -> 9.8 | Helped |
| Add mean teacher (ablation) | ConvMLP-Pin detector with DFDN, FSH, and hard mixup · City Persons | MR−2: 9.8 -> 8.7 | Helped |
| Add SP3 (ablation) | ConvMLP-Pin detector with DFDN, FSH, hard mixup, and mean teacher · City Persons | MR−2: 8.7 -> 8.7 | No clear effect |
| Use LSFM instead of F2DNet | Pedestrian detection · Euro City Persons reasonable | MR−2: 6.1 -> 4.7 | Helped |
| Use LSFM instead of F2DNet | Pedestrian detection · Caltech reasonable | MR−2: 2.2 -> 3.1 | Hurt |
| Use LSFM instead of F2DNet | Pedestrian detection · City Persons reasonable | MR−2: 8.7 -> 8.5 | Helped |
Closest papers by meaning
- Generalizable Pedestrian Detection: The Elephant in the Room (CVPR 2021 · similarity 0.88)
- A Hybrid Global-Local Perception Network for Lane Detection (AAAI 2024 · similarity 0.85)
- Weakly Supervised Semantic Segmentation for Driving Scenes (AAAI 2024 · similarity 0.85)
- Multi-View Pedestrian Occupancy Prediction with a Novel Synthetic Dataset (AAAI 2025 · similarity 0.85)
- GSENet:Global Semantic Enhancement Network for Lane Detection (AAAI 2024 · similarity 0.84)
- Roboflow100-VL: A Multi-Domain Object Detection Benchmark for Vision-Language Models (NeurIPS 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.