Research map / 3D vision / Face and shape
Learning Adversarial 3D Model Generation With 2D Image Enhancer
Read the paperCheck 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 |
|---|---|---|---|
| add image enhancer feeding learned image features to 3D generator (ablation) | 3D-GAN-style generator · ModelNet10 classification | accuracy: 88.88 -> 91.63 | Helped |
| add image enhancer (ablation) | 3D-GAN-style generator · ModelNet40 classification | accuracy: 85.53 -> 87.85 | Helped |
| add image enhancer (ablation) | 3D-GAN-style generator · ModelNet10 retrieval | MAP: 61.82 -> 65.00 | Helped |
| add image enhancer (ablation) | 3D-GAN-style generator · ModelNet40 retrieval | MAP: 40.81 -> 44.44 | Helped |
| feed enhancer features from deeper layer (2nd, 3rd) instead of 1st (ablation) | enhancer-driven 3D generator · ModelNet10 classification | accuracy: 91.63 -> 91.52 (2nd), 90.42 (3rd) | Hurt |
| feed enhancer features from deeper layer (2nd, 3rd) instead of 1st (ablation) | enhancer-driven 3D generator · ModelNet40 classification | accuracy: 87.85 -> 87.12 (2nd), 86.34 (3rd) | Hurt |
| use enhancer framework instead of 3D-GAN reported result | unsupervised shape features · ModelNet10 classification | accuracy: 91.00 -> 91.63 | Helped |
| use enhancer framework instead of 3D-GAN reported result | unsupervised shape features · ModelNet40 classification | accuracy: 83.30 -> 87.85 | Helped |
Closest papers by meaning
- Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling (NeurIPS 2016 · similarity 0.91)
- Efficient Geometry-Aware 3D Generative Adversarial Networks (CVPR 2022 · similarity 0.88)
- What You See is What You GAN: Rendering Every Pixel for High-Fidelity Geometry in 3D GANs (CVPR 2024 · similarity 0.88)
- 3D Shape Generation With Grid-Based Implicit Functions (CVPR 2021 · similarity 0.88)
- 3D generation on ImageNet (ICLR 2023 · similarity 0.88)
- Rank3DGAN: Semantic Mesh Generation Using Relative Attributes (AAAI 2020 · 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.