Research map / Point clouds / Point cloud classification
Point-Focused Attention Meets Context-Scan State Space: Robust Biological Visual Perception for Point Cloud Representation
Kanglin Qu, Pan Gao, Qun Dai, Yuanhao Sun
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 local neighbor branch to point-focused attention (ablation) | PointLearner · ModelNet40 | OA: 92.11 -> 94.17 | Improved |
| add spatial downsampling branch (ablation) | PointLearner · ModelNet40 | OA: 93.06 -> 94.17 | Improved |
| competitive single-softmax fusion instead of additive fusion (ablation) | PointLearner · ModelNet40 | OA: 93.43 -> 94.17 | Improved |
| bidirectional S6 instead of unidirectional S6 (ablation) | PointLearner · ModelNet40 | OA: 93.08 -> 94.17 | Improved |
| add context-scan state space on top of point-focused attention (ablation) | PointLearner · ModelNet40 | OA: 92.93 -> 94.17 | Improved |
| add point-focused attention to context-scan state space (ablation) | PointLearner · ModelNet40 | OA: 91.94 -> 94.17 | Improved |
| Hilbert serialization instead of none (ablation) | PointLearner · ModelNet40 | OA: 91.34 -> 94.17 | Improved |
| Hilbert instead of Z-Order curve (ablation) | PointLearner · ModelNet40 | OA: 93.06 -> 94.17 | Improved |
| combine two serialization curves instead of Hilbert alone (ablation) | PointLearner · ModelNet40 | OA: 94.17 -> 93.78 (Hilbert & Trans-Hilbert), 93.52 (Hilbert & Z-Order) | Hurt |
| learnable serialization instead of Hilbert curve (ablation) | PointLearner · ModelNet40 | OA: 94.17 -> 92.78 | Hurt |
| use induced point pooling instead of FPS for spatial downsampling at high rate (/256) (ablation) | PointLearner · ModelNet40 | OA: IPP 93.39 at /256; FPS approaches no-branch baseline | 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.
- PointMamba — PointMamba: A Simple State Space Model for Point Cloud Analysis (NeurIPS 2024) · better on 1 of 1 shared result
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
- Into the Rabbit Hull: From Task-Relevant Concepts in DINO to Minkowski Geometry (ICLR 2026 · similarity 0.86)
- Foveated Reasoning: Stateful, Action-based Visual Focusing for Vision-Language Models (ECCV 2026 · similarity 0.86)
- Attentional ShapeContextNet for Point Cloud Recognition (CVPR 2018 · similarity 0.86)
- Clustering-based Supervised Training for 3D Perception (TPAMI 2026 · similarity 0.86)
- Sonata: Self-Supervised Learning of Reliable Point Representations (CVPR 2025 · similarity 0.85)
- TransNeXt: Robust Foveal Visual Perception for Vision Transformers (CVPR 2024 · similarity 0.85)
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.
