Research map / Point clouds / Point cloud classification
Self-Supervised Learning on 3D Point Clouds by Learning Discrete Generative Models
Check 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 |
|---|---|---|---|
| Pre-train on ShapeNet with ParAE before fitting a linear SVM | DGCNN · ModelNet40 test | Accuracy%: 90.8 -> 91.6 | Helped |
| Pre-train on ShapeNet with ParAE before fitting a linear SVM | PointNet · ModelNet40 test | Accuracy%: 88.7 -> 90.3 | Helped |
| Initialize from ParAE pre-training rather than random weights | PointNet · ModelNet40 test | Accuracy: 89.2 -> 90.5 | Helped |
| Initialize from ParAE pre-training rather than random weights | DGCNN · ModelNet40 test | Accuracy: 92.2 -> 92.9 | Helped |
| Use ParAE pre-training rather than OcCo pre-training | PointNet · ModelNet40 test | Accuracy: 90.2 -> 90.5 | Helped |
| Use ParAE pre-training rather than OcCo pre-training | DGCNN · ModelNet40 test | Accuracy: 93.1 -> 92.9 | Hurt |
| Pre-train on Areas 1-4 and 6 with ParAE before supervised fine-tuning | DGCNN · S3DIS Area 5 test (train Area 1) | mIOU: +9.9 | Helped |
| Pre-train on Areas 1-4 and 6 with ParAE before supervised fine-tuning | DGCNN · S3DIS Area 5 test (train Area 2) | mIOU: +3.9 | Helped |
| Pre-train on Areas 1-4 and 6 with ParAE before supervised fine-tuning | DGCNN · S3DIS Area 5 test (train Area 3) | mIOU: +8.5 | Helped |
| Pre-train on Areas 1-4 and 6 with ParAE before supervised fine-tuning | DGCNN · S3DIS Area 5 test (train Area 4) | mIOU: +5.6 | Helped |
| Pre-train on Areas 1-4 and 6 with ParAE before supervised fine-tuning | DGCNN · S3DIS Area 5 test (train Area 6) | mIOU: +5.3 | Helped |
| Increase the number of inferred spatial partitions J (ablation) | ParAE · Downstream tasks | Performance | Helped |
Closest papers by meaning
- DOS: Distilling Observable Softmaps of Zipfian Prototypes for Self-Supervised Point Representation (AAAI 2026 · similarity 0.90)
- Sonata: Self-Supervised Learning of Reliable Point Representations (CVPR 2025 · similarity 0.89)
- Self-Supervised Deep Learning on Point Clouds by Reconstructing Space (NeurIPS 2019 · similarity 0.89)
- Multi-View Representation is What You Need for Point-Cloud Pre-Training (ICLR 2024 · similarity 0.88)
- X-3D: Explicit 3D Structure Modeling for Point Cloud Recognition (CVPR 2024 · similarity 0.88)
- FreePoint: Unsupervised Point Cloud Instance Segmentation (CVPR 2024 · similarity 0.87)
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.