Research map / Point clouds / Segmentation and detection
Multimodal and open-vocabulary 3D detection
Improve 3D detection with transformers, vision fusion, multi-domain training and open-vocabulary labels.
24 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2019–2025) · an approach in Segmentation and detection, keyword cluster “detection · 3d object · detectors”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
10 papers in 2023–24 and 6 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 10.0% to 6.6%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2019 · PointPillars: Fast Encoders for Object Detection From Point Clouds (CVPR 2019 · 4,587 citations)
- 2020 · Associate-3Ddet: Perceptual-to-Conceptual Association for 3D Point Cloud Object Detection (CVPR 2020 · 106 citations)
- 2021 · CIA-SSD: Confident IoU-Aware Single-Stage Object Detector From Point Cloud (AAAI 2021 · 291 citations)
- 2022 · RBGNet: Ray-Based Grouping for 3D Object Detection (CVPR 2022 · 70 citations)
- 2023 · PointDistiller: Structured Knowledge Distillation Towards Efficient and Compact 3D Detection (CVPR 2023 · 63 citations)
- 2024 · Voxel Mamba: Group-Free State Space Models for Point Cloud based 3D Object Detection (NeurIPS 2024 · 35 citations)
- 2025 · SpatialLM: Training Large Language Models for Structured Indoor Modeling (NeurIPS 2025 · 8 citations)
Earliest papers here
- PointPillars: Fast Encoders for Object Detection From Point Clouds (CVPR 2019 · 4,587 citations)
- Associate-3Ddet: Perceptual-to-Conceptual Association for 3D Point Cloud Object Detection (CVPR 2020 · 106 citations)
- CIA-SSD: Confident IoU-Aware Single-Stage Object Detector From Point Cloud (AAAI 2021 · 291 citations)
What has been tried
23 papers here report 269 measured changes (a component added, removed or swapped, compared with a baseline): helped 81%, no clear effect 4%, hurt 14%, mixed 1%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use Hilbert Curve instead of random serialization | nuScenes validation · mAP | 66.0 -> 67.5 | Helped | Voxel Mamba: Group-Free State Space Models for Point Cloud based 3D Object Detection (NeurIPS 2024 · 35 citations) |
| Use Sonata/PTv3 instead of the mapping-based or 3DCNN encoder alternatives | Encoder design experiment · Layout F1 IoU2D@0.25 | 10.8 -> 84.6 | Helped | SpatialLM: Training Large Language Models for Structured Indoor Modeling (NeurIPS 2025 · 8 citations) |
| Add scale-aligned fusion (SAF) | KITTI val, Car easy · 3D AP | 92.38 -> 92.93 | Helped | SSLFusion: Scale and Space Aligned Latent Fusion Model for Multimodal 3D Object Detection (AAAI 2025 · 6 citations) |
| replace D-FPS with active sampling in layers 3–4 | KITTI val, three-class · APR40 mean | 69.50 -> 73.83 | Helped | AS-Det: Active Sampling for Adaptive 3D Object Detection in Point Clouds (AAAI 2025 · 5 citations) |
| Add grid search to Max-min calibration | Waymo val (L2) · Mean mAPH | 52.91 -> 58.39 | Helped | LiDAR-PTQ: Post-Training Quantization for Point Cloud 3D Object Detection (ICLR 2024 · 4 citations) |
| add train-phase prior size filtering | SUNRGBD · mAP@0.25 | +1.65% | Helped | ImOV3D: Learning Open Vocabulary Point Clouds 3D Object Detection from Only 2D Images (NeurIPS 2024 · 3 citations) |
| Add Inner Mamba instead of using PointNet++ for local encoding | ScanNet val · mAP@0.25 | 63.1 -> 64.9 | Helped | 3DET-Mamba: Causal Sequence Modelling for End-to-End 3D Object Detection (NeurIPS 2024 · 3 citations) |
| use all three views (XY+YZ+XZ) instead of only XY | PCSOD · IoU | 0.719 -> 0.763 | Helped | MSV-PCT: Multi-Sparse-View Enhanced Transformer Framework for Salient Object Detection in Point Clouds (AAAI 2025 · 2 citations) |
| domain-aware scatter partitioning (per-domain normalization scale/shift) | SUN RGB-D + KITTI joint · AP_25 SUN RGB-D | 61.9 -> 63.2 | Helped | One for All: Multi-Domain Joint Training for Point Cloud Based 3D Object Detection (NeurIPS 2024 · 1 citation) |
| Use multiple Hilbert axis orders across decoder layers instead of one order | ScanNet V2 · AP50 | 51.1 -> 52.7 | Helped | State Space Model Meets Transformer: A New Paradigm for 3D Object Detection (ICLR 2025) |
What reviewers will expect
Datasets and baselines named in the 6 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- KITTI · 2 of 6 papers
Baselines
- SECOND · 2 of 6 papers
- PointRCNN · 2 of 6 papers
- V-DETR · 2 of 6 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 16 papers; from OpenAlex and the papers' affiliations).
- Tsinghua University · 3 papers
- The University of Hong Kong · 2 papers
- Fudan University · 2 papers
- The Hong Kong Polytechnic University · 2 papers
- Southeast University · 2 papers
- Chinese Academy of Sciences · 2 papers
Use it
Download BibTeX of 24 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- LiDAR voxel and pillar detectors (56 papers)
- Single-stage point-voxel detection (45 papers)
Elsewhere in the topic
- Point cloud semantic segmentation (108 papers)
- LiDAR segmentation for driving (47 papers)
- Scene completion and instance segmentation (44 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).