Research map / 3D vision / Face and shape
Generative VoxelNet: Learning Energy-Based Models for 3D Shape Synthesis and Analysis
Jianwen Xie, Zilong Zheng, Ruiqi Gao, Wenguan Wang, Song Chun Zhu, Ying Nian Wu
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 |
|---|---|---|---|
| replace Primitive GAN with generative VoxelNet | Primitive GAN · ModelNet10 synthesis | Inception score: 11.520 -> 11.772 | Improved |
| replace 3D GAN with generative VoxelNet | 3D GAN · ModelNet10 synthesis | Inception score: 8.658 -> 11.772 | Improved |
| replace best baseline with generative VoxelNet | baseline 1 · ModelNet10 single-category | softmax class probability: 0.7431 -> 0.8811 | Improved |
| replace best baseline with generative VoxelNet | baseline 1 · ModelNet10 single-category | classification error: 0.2315 -> 0.0968 | Improved |
| use 3-layer instead of 2-layer conditional model for recovery (ablation) | generative VoxelNet · ModelNet10 recovery | recovery error: 0.0127 -> 0.0092 | Improved |
| replace baseline recovery with conditional generative VoxelNet | baseline 1 · ModelNet10 recovery | recovery error: 0.0259 -> 0.0092 | Improved |
| replace Primitive GAN features with generative VoxelNet features | Primitive GAN · ModelNet10 classification | accuracy: 92.2% -> 92.4% | Improved |
| replace VoxNet with generative VoxelNet unsupervised features | VoxNet · ModelNet10 classification | accuracy: 92.0% -> 92.4% | No clear effect |
Closest papers by meaning
- Learning Descriptor Networks for 3D Shape Synthesis and Analysis (CVPR 2018 · similarity 0.91)
- Synthesizing 3D Shapes via Modeling Multi-View Depth Maps and Silhouettes With Deep Generative Networks (CVPR 2017 · similarity 0.89)
- 3D Shape Generation and Completion Through Point-Voxel Diffusion (ICCV 2021 · similarity 0.88)
- Deep Optimized Priors for 3D Shape Modeling and Reconstruction (CVPR 2021 · similarity 0.88)
- ShapeShifter: 3D Variations Using Multiscale and Sparse Point-Voxel Diffusion (CVPR 2025 · similarity 0.88)
- Make-A-Shape: a Ten-Million-scale 3D Shape Model (ICML 2024 · similarity 0.87)
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.
