Research map / Point clouds / Segmentation and detection
MemorySeg: Online LiDAR Semantic Segmentation with a Latent Memory
Enxu Li, Sergio Casas, Raquel Urtasun
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 sparse 3D latent memory (FAM + MRM) to single-frame baseline (ablation) | MemorySeg · SemanticKITTI val | mIoU (%): 67.2 -> 69.5 | Improved |
| add adaptive padding module (APM) on top of memory (ablation) | MemorySeg · SemanticKITTI val | mIoU (%): 69.5 -> 70.8 | Improved |
| use memory refinement module instead of vanilla sparse ConvGRU (ablation) | MemorySeg · SemanticKITTI val | mIoU (%): 69.7 -> 70.8 | Improved |
| add variation regularizer (ablation) | MemorySeg · nuScenes val | mIoU (%): 72.9 -> 76.7 | Improved |
| add variation regularizer (ablation) | MemorySeg · SemanticKITTI val | mIoU (%): 66.4 -> 67.2 | Improved |
| add variation regularizer (ablation) | MemorySeg · PandaSet val | mIoU (%): 64.7 -> 66.6 | Improved |
| add latent memory, evaluated by distance range | single frame baseline · nuScenes val long range | mIoU | Improved |
| use 3D latent memory rather than range-view latent memory | TemporalLidarSeg vs MemorySeg · SemanticKITTI multi-scan test | mIoU (%): 47.0 -> 58.3 | Improved |
| replace memory with naive concatenation of past five frames (ablation) | MemorySeg · runtime | runtime: 1.58 times longer than SFB | Hurt |
Closest papers by meaning
- LESS: Label-Efficient Semantic Segmentation for LiDAR Point Clouds (ECCV 2022 · similarity 0.92)
- TASeg: Temporal Aggregation Network for LiDAR Semantic Segmentation (CVPR 2024 · similarity 0.92)
- Less Is More: Reducing Task and Model Complexity for 3D Point Cloud Semantic Segmentation (CVPR 2023 · similarity 0.91)
- LaserMix for Semi-Supervised LiDAR Semantic Segmentation (CVPR 2023 · similarity 0.91)
- TREND: Unsupervised 3D Representation Learning via Temporal Forecasting for LiDAR Perception (NeurIPS 2025 · similarity 0.90)
- PC-BEV: An Efficient Polar-Cartesian BEV Fusion Framework for LiDAR Semantic Segmentation (AAAI 2025 · similarity 0.89)
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.
