Research map / Point clouds / Point cloud classification
Neural Implicit Embedding for Point Cloud Analysis
Kent Fujiwara, Taiichi Hashimoto
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 |
|---|---|---|---|
| Concatenate point coordinates and normals to the ELM-weight representation (ablation) | Neural Implicit Embedding · ModelNet 10 | classification accuracy (%): 95.7 -> 96.7 | Improved |
| Concatenate point coordinates and normals to the ELM-weight representation (ablation) | Neural Implicit Embedding · ModelNet 40 | classification accuracy (%): 92.2 -> 93.2 | Improved |
| Reduce the original point-cloud size from 2048 to 1024 points (ablation) | Neural Implicit Embedding · ModelNet 40 | classification accuracy (%): 93.2 -> 92.7 | Hurt |
| Reduce the original point-cloud size from 2048 to 512 points (ablation) | Neural Implicit Embedding · ModelNet 40 | classification accuracy (%): 93.2 -> 91.1 | Hurt |
| Reduce the original point-cloud size from 2048 to 256 points (ablation) | Neural Implicit Embedding · ModelNet 40 | classification accuracy (%): 93.2 -> 90.1 | Hurt |
| Add noise to the representation (ablation) | Neural Implicit Embedding · ModelNet classification | classification accuracy | Hurt |
| Use canonical projection to align distance fields under rotation | Neural Implicit Embedding · ModelNet 40 with test-time rotation perturbations | classification accuracy (%): 93.2 -> 81.8 at maximum rotation angle π | Improved |
| Use the scale-invariant representation under rotation and scale perturbation | Neural Implicit Embedding · ModelNet 40, rotation and scale perturbation | classification accuracy (%): 83.9 | Improved |
| Use Neural Implicit Embedding instead of PointNet++ for ShapeNetPart segmentation | Point-cloud segmentation · ShapeNetPart | mean IoU: 85.1 -> 85.2 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- SO-Net — SO-Net: Self-Organizing Network for Point Cloud Analysis (CVPR 2018) · better on 1 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
- Point Deformable Network with Enhanced Normal Embedding for Point Cloud Analysis (AAAI 2024 · similarity 0.88)
- Clustering based Point Cloud Representation Learning for 3D Analysis (ICCV 2023 · similarity 0.87)
- Neural Intrinsic Embedding for Non-Rigid Point Cloud Matching (CVPR 2023 · similarity 0.87)
- Self-Supervised Arbitrary-Scale Point Clouds Upsampling via Implicit Neural Representation (CVPR 2022 · similarity 0.87)
- Learning Continuous Implicit Field with Local Distance Indicator for Arbitrary-Scale Point Cloud Upsampling (AAAI 2024 · similarity 0.87)
- LinNet: Linear Network for Efficient Point Cloud Representation Learning (NeurIPS 2024 · 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.
