Research map / 3D vision / Face and shape
RIM-Net: Recursive Implicit Fields for Unsupervised Learning of Hierarchical Shape Structures
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 |
|---|---|---|---|
| Use per-point Gaussian decoding instead of a scalar occupancy output per point (ablation) | RIM-Net hierarchical shape reconstruction · ShapeNet categories (ablation mean) | per-label IoU: 65.1 -> 80.2 | Helped |
| Use per-point Gaussian decoding instead of a scalar occupancy output per point (ablation) | RIM-Net hierarchical shape reconstruction · ShapeNet categories (ablation mean) | CD: 0.4909 -> 0.4605 | Helped |
| Add the decomposition loss enforcing that each parent is the union of its children (ablation) | RIM-Net hierarchy · ShapeNet categories (ablation mean) | per-label IoU: 75.5 -> 80.2 | Helped |
| Add progressive level-wise training followed by fine-tuning (ablation) | RIM-Net hierarchy · ShapeNet categories (ablation mean) | CD: 0.8448 -> 0.4605 | Helped |
| Add progressive level-wise training followed by fine-tuning (ablation) | RIM-Net hierarchy · ShapeNet categories (ablation mean) | per-label IoU: 67.6 -> 80.2 | Helped |
| Use RIM-Net rather than the strongest reported baseline, HA | Unsupervised structure reconstruction · ShapeNet airplane, chair, table (mean) | CD: 0.6678 -> 0.4605 | Helped |
| Use RIM-Net rather than IM-Net | Single-view shape reconstruction · ShapeNet categories (mean) | CD: 0.6575 -> 0.6291 | Helped |
| Use RIM-Net rather than IM-Net | Single-view shape reconstruction · ShapeNet categories (mean) | LFD: 4388.74 -> 4273.19 | Helped |
| Use RIM-Net rather than IM-Net for single-view chair reconstruction | Single-view reconstruction · Chair | LFD: 3499.15 -> 3600.60 | Hurt |
| Use RIM-Net rather than IM-Net for single-view table reconstruction | Single-view reconstruction · Table | CD: 0.8762 -> 1.1223 | Hurt |
Closest papers by meaning
- PartNet: A Recursive Part Decomposition Network for Fine-Grained and Hierarchical Shape Segmentation (CVPR 2019 · similarity 0.91)
- Learning Unsupervised Hierarchical Part Decomposition of 3D Objects From a Single RGB Image (CVPR 2020 · similarity 0.90)
- Neural Parts: Learning Expressive 3D Shape Abstractions With Invertible Neural Networks (CVPR 2021 · similarity 0.89)
- Im2Struct: Recovering 3D Shape Structure From a Single RGB Image (CVPR 2018 · similarity 0.89)
- StructEdit: Learning Structural Shape Variations (CVPR 2020 · similarity 0.89)
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.88)
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.