Research map / 3D vision / Face and shape
SkeletonNet: A Topology-Preserving Solution for Learning Mesh Reconstruction of Object Surfaces From RGB Images
Jiapeng Tang, Xiaoguang Han, Mingkui Tan, Xin Tong, Kui Jia
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 Laplacian regularization to skeletal point regression (ablation) | SkeletonNet · ShapeNet chair | Chamfer distance x0.001: 1.911 -> 1.485 | Improved |
| add Laplacian regularization (ablation) | SkeletonNet · ShapeNet chair | IoU x100: 31.10 -> 39.50 | Improved |
| add end-to-end training (ablation) | SkeletonNet · ShapeNet chair | Chamfer distance x0.001: 1.601 -> 1.485 | Improved |
| add end-to-end training (ablation) | SkeletonNet · ShapeNet chair | IoU x100: 37.73 -> 39.50 | Improved |
| use curve and sheet primitive decoders instead of point-wise fitting (ablation) | SkeletonNet · ShapeNet chair | IoU x100: 28.97 -> 39.50 | Improved |
| use separate curve and sheet decoders instead of shared line-and-square fitting (ablation) | SkeletonNet · ShapeNet chair | Chamfer distance x0.001: 2.451 -> 1.485 | Improved |
| use globally guided subvolume synthesis instead of subvolume synthesis alone (ablation) | SkeletonNet · ShapeNet chair | IoU x100: 37.64 -> 39.50 | Improved |
| use subvolume synthesis instead of quantization and morphological dilation (ablation) | SkeletonNet · ShapeNet chair | IoU x100: 27.52 -> 39.50 | Improved |
| use SkeletonNet skeletal volumes instead of OGN-synthesized volumes (ablation) | SkeGCNN · ShapeNet chair | Chamfer distance x0.001: 4.485 -> 1.154 | Improved |
| use SkeletonNet skeletal volumes instead of OGN-synthesized volumes (ablation) | SkeGCNN · ShapeNet chair | IoU x100: 49.84 -> 62.90 | Improved |
| use SkeletonNet skeletal volumes instead of OGN-synthesized volumes (ablation) | SkeDISN · ShapeNet chair | Chamfer distance x0.001: 1.534 -> 1.230 | Improved |
| use SkeletonNet skeletal volumes instead of OGN-synthesized volumes (ablation) | SkeDISN · ShapeNet chair | IoU x100: 61.70 -> 65.20 | Improved |
Closest papers by meaning
- A Skeleton-Bridged Deep Learning Approach for Generating Meshes of Complex Topologies From Single RGB Images (CVPR 2019 · similarity 0.95)
- Point2Skeleton: Learning Skeletal Representations from Point Clouds (CVPR 2021 · similarity 0.90)
- Skeleton-bridged Point Completion: From Global Inference to Local Adjustment (NeurIPS 2020 · similarity 0.89)
- Voxurf: Voxel-based Efficient and Accurate Neural Surface Reconstruction (ICLR 2023 · similarity 0.88)
- Learning Unsupervised Hierarchical Part Decomposition of 3D Objects From a Single RGB Image (CVPR 2020 · similarity 0.88)
- Learning Implicit Representation for Reconstructing Articulated Objects (ICLR 2024 · 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.
