Research map / Point clouds / Registration and matching
FS-I2P: A Hierarchical Focus–Sweep Registration Network with Dynamically Allocated Depth
Zhixin Cheng, Yujia Chen, Xujing Tao, Bohao Liao, Xiaotian Yin, Baoqun Yin, Tianzhu Zhang
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 Sweep module to HTS baseline (ablation) | FS-I2P · RGB-D Scenes V2 | RR: 56.4 -> 62.4 | Improved |
| add Focus on top of Sweep + HTS (ablation) | FS-I2P · RGB-D Scenes V2 | RR: 62.4 -> 68.2 | Improved |
| add Focus without Sweep (ablation) | FS-I2P · RGB-D Scenes V2 | RR: 56.4 -> 59.7 | Improved |
| add Dynamic Layer Allocation Strategy (ablation) | FS-I2P · RGB-D Scenes V2 | RR: 68.2 -> 72.9 | Improved |
| add hierarchical top-k selection (ablation) | FS-I2P · RGB-D Scenes V2 | RR: 56.2 -> 72.9 | Improved |
| remove all modules including HTS (ablation) | FS-I2P · RGB-D Scenes V2 | RR: 56.4 -> 49.5 | Hurt |
| stack more transformer cross-attention layers (ablation) | Transformer interaction · image-point MMD | MMD: improves then rises with depth | Mixed |
| use DINOv2 image encoder | FS-I2P · RGB-D Scenes V2 | RR: 72.9 -> 86.3 | Improved |
| use DINOv2 image encoder | FS-I2P · 7-Scenes | RR: 82.7 -> 84.6 | Improved |
| use Mamba-based Focus-Sweep instead of transformer interaction | FS-I2P · efficiency | inference time and memory: 0.497 s -> 0.304 s; 8346 -> 6415 MB | Improved |
| use FS-I2P instead of Diff2I2P on 7-Scenes | FS-I2P · 7-Scenes | RR: 83.0 vs 82.7 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- 2D3D-MATR — 2D3D-MATR: 2D-3D Matching Transformer for Detection-Free Registration Between Images and Point Clouds (ICCV 2023) · better on 7 of 7 shared results
- Diff2I2P — Diff2I2P: Differentiable Image-to-Point Cloud Registration with Diffusion Prior (ICCV 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
- Dual Focus-Attention Transformer for Robust Point Cloud Registration (CVPR 2025 · similarity 0.88)
- 3D Focusing-and-Matching Network for Multi-Instance Point Cloud Registration (NeurIPS 2024 · similarity 0.88)
- Dynamic Cues-Assisted Transformer for Robust Point Cloud Registration (CVPR 2024 · similarity 0.88)
- Bridge 2D-3D: Uncertainty-aware Hierarchical Registration Network with Domain Alignment (AAAI 2025 · similarity 0.87)
- Generative Point Cloud Registration (ICML 2025 · similarity 0.86)
- FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimators (ICLR 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.
