Research map / Detection and segmentation / 3D detection and driving
Are We Ready for Vision-Centric Driving Streaming Perception? The ASAP Benchmark
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 |
|---|---|---|---|
| Add velocity-based updating with Kalman filtering | BEVDepth · nuScenes-H val, RTX3090 | mAP-S: 0.323 -> 0.328 | Helped |
| Add velocity-based updating with Kalman filtering | BEVDepth · nuScenes-H val, RTX2070S | mAP-S: 0.306 -> 0.316 | Helped |
| Add velocity-based updating with Kalman filtering | BEVDepth · nuScenes-H val, GTX1060 | mAP-S: 0.226 -> 0.263 | Helped |
| Forecast next-frame object locations | BEVDepth · nuScenes-H val, RTX3090 | mAP-S: 0.323 -> 0.329 | Helped |
| Forecast next-frame object locations | BEVDepth · nuScenes-H val, RTX2070S and GTX1060 | mAP-S: 3.0%, 4.0% | Mixed |
| Add constant-velocity motion updating (ablation) | BEVFormer · nuScenes-H val, RTX3090 | mAP-S: 0.310 -> 0.332 | Helped |
| Add Kalman-filter refinement to constant-velocity updating (ablation) | BEVFormer · nuScenes-H val, RTX3090 | mAP-S: 0.332 -> 0.344 | Helped |
| Add constant-velocity motion updating (ablation) | FCOS3D · nuScenes-H val, RTX3090 | mAP-S: 0.208 -> 0.212 | Helped |
| Add Kalman-filter refinement to constant-velocity updating (ablation) | FCOS3D · nuScenes-H val, RTX3090 | mAP-S: 0.212 -> 0.218 | Helped |
| Replace ResNet-50 with ResNet-101 (ablation) | BEVDepth, 704×256 input · nuScenes-H val, RTX3090 | mAP-S: 0.323 -> 0.331 | Helped |
| Increase input size from 704×256 to 1408×512 (ablation) | BEVDepth R-101 · nuScenes-H val, RTX3090 | mAP-S: 0.331 -> 0.311 | Hurt |
| Increase input size from 704×256 to 1408×512 (ablation) | BEVDepth-Sv R-101 · nuScenes-H val, RTX3090 | mAP-S: 0.340 -> 0.341 | Helped |
Closest papers by meaning
- Real-time Stereo-based 3D Object Detection for Streaming Perception (NeurIPS 2024 · similarity 0.88)
- OpenAD: Open-World Autonomous Driving Benchmark for 3D Object Detection (NeurIPS 2025 · similarity 0.86)
- UniDrive: Towards Universal Driving Perception Across Camera Configurations (ICLR 2025 · similarity 0.86)
- Transtreaming: Adaptive Delay-aware Transformer for Real-time Streaming Perception (AAAI 2025 · similarity 0.86)
- VisionPAD: A Vision-Centric Pre-training Paradigm for Autonomous Driving (CVPR 2025 · similarity 0.85)
- Glad: A Streaming Scene Generator for Autonomous Driving (ICLR 2025 · similarity 0.85)
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.