Research map / Point clouds / Point cloud classification
Utonia: Toward One Encoder for All Point Clouds
Yujia Zhang, Xiaoyang Wu, Yunhan Yang, Xianzhe Fan, Han Li, Yuechen Zhang, Zehao Huang, Naiyan Wang, Hengshuang Zhao
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 RoPE on granularity-aligned coordinates (ablation) | Utonia multi-data · ScanNet200 / Waymo / PartNetE | mIoU: 33.5 -> 34.9; 56.6 -> 59.2; 43.5 -> 44.6 | Improved |
| add RoPE in single-domain training (ablation) | Utonia single-data · ScanNet200 / PartNetE | mIoU: 34.4 -> 34.0; 44.1 -> 42.4 | Hurt |
| add RoPE in single-domain training (ablation) | Utonia single-data · Waymo | mIoU: 60.5 -> 62.1 | Improved |
| vary RoPE base from 1 to 1000 (ablation) | Utonia · ScanNet200 / Waymo / PartNetE | mIoU: 34.0, 34.9, 34.1, 34.9 on ScanNet200 | No clear effect |
| add modality blinding at data loading (ablation) | Utonia · ScanNet200 w/o color | mIoU: 9.3 -> 31.7 | Improved |
| drop modality on masked views instead of at loading (ablation) | Utonia · ScanNet200 w/o color | mIoU: 31.7 -> 10.1 | Hurt |
| use full-strength object scale augmentation (vs less scale aug.) (ablation) | Utonia · ScanObjectNN(H) | mIoU: 64.5 -> 66.9 | Improved |
| use strong object rotation augmentation (vs less rot. aug.) (ablation) | Utonia · ScanObjectNN(H) | mIoU: 63.0 -> 66.9 | Improved |
| use strong rotation augmentation for objects (vs less rot. aug.) (ablation) | Utonia · PartNetE | mIoU: 47.5 -> 44.6 | Hurt |
| add temporal frame augmentation (ablation) | Utonia · PartNetE / Waymo | mIoU: 40.0 -> 44.6; 58.8 -> 59.2 | Improved |
| add RoPE coordinate augmentation (ablation) | Utonia · Waymo | mIoU: 60.1 -> 59.2 | No clear effect |
| scale model from 38M to 137M with 83k data (ablation) | Utonia · Waymo linear probing | mIoU: 59.2 -> 62.6 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Sonata — Sonata: Self-Supervised Learning of Reliable Point Representations (CVPR 2025) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Invaria: Learning Scale and Density Invariance in Point Clouds via Next-Resolution Prediction (NeurIPS 2026 · similarity 0.87)
- Point Cloud Matters: Rethinking the Impact of Different Observation Spaces on Robot Learning (NeurIPS 2024 · similarity 0.87)
- Sonata: Self-Supervised Learning of Reliable Point Representations (CVPR 2025 · similarity 0.86)
- PointMAC: Meta-Learned Adaptation for Robust Test-Time Point Cloud Completion (NeurIPS 2025 · similarity 0.86)
- Adapting Pre-trained 3D Models for Point Cloud Video Understanding via Cross-frame Spatio-temporal Perception (CVPR 2025 · similarity 0.86)
- PointMapPolicy: Structured Point Cloud Processing for Multi-Modal Imitation Learning (NeurIPS 2025 · similarity 0.86)
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.
