Research map / 3D vision / 3D generation
Texture Generation Using Dual-Domain Feature Flow with Multi-View Hallucinations
Seunggyu Chang, Jungchan Cho, Songhwai Oh
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 |
|---|---|---|---|
| Use the full dual-domain method rather than the strongest listed baseline | Pose transfer and texture-rendering system · DeepFashion | FID: 10.573 -> 9.001 | Improved |
| Use the full dual-domain method rather than the strongest listed baseline | Pose transfer and texture-rendering system · DeepFashion | LPIPS: 0.234 -> 0.156 | Improved |
| Use the full dual-domain method rather than the strongest listed baseline | Pose transfer and texture-rendering system · DeepFashion | SSIM: 0.763 -> 0.830 | Improved |
| Use the full dual-domain method rather than the strongest listed baseline | Pose transfer and texture-rendering system · iPER | LPIPS: 0.087 -> 0.051 | Improved |
| Use the full dual-domain method rather than the strongest listed baseline | Pose transfer and texture-rendering system · iPER | SSIM: 0.854 -> 0.907 | Improved |
| Add image-to-texture attention to the independent model (ablation) | Texture generator G_T · iPER | LPIPS: 0.075 -> 0.073 | Improved |
| Add image-to-texture attention to the independent model (ablation) | Texture generator G_T · FashionVid | LPIPS: 0.100 -> 0.083 | Improved |
| Add texture-to-image attention to the image-to-texture-only model (ablation) | Texture generator G_T · iPER | LPIPS: 0.073 -> 0.063 | Improved |
| Add texture-to-image attention to the image-to-texture-only model (ablation) | Texture generator G_T · FashionVid | SSIM: 0.902 -> 0.903 | Improved |
| Add both inter-domain attention paths (ablation) | Image generator G_I · FashionVid | SSIM: 0.913 -> 0.921 | Improved |
| Add either single inter-domain attention path (ablation) | Image generator G_I · iPER | LPIPS: 0.051 -> 0.056; 0.051 -> 0.061 | Mixed |
| Increase hallucination views from one to two (ablation) | G_I and G_T · iPER | SSIM: 0.894 -> 0.907; 0.887 -> 0.898 | Improved |
Closest papers by meaning
- Hallo3D: Multi-Modal Hallucination Detection and Mitigation for Consistent 3D Content Generation (NeurIPS 2024 · similarity 0.88)
- Paint3D: Paint Anything 3D with Lighting-Less Texture Diffusion Models (CVPR 2024 · similarity 0.86)
- ViewDiff: 3D-Consistent Image Generation with Text-to-Image Models (CVPR 2024 · similarity 0.85)
- DreamCraft3D: Hierarchical 3D Generation with Bootstrapped Diffusion Prior (ICLR 2024 · similarity 0.84)
- MeshGen: Generating PBR Textured Mesh with Render-Enhanced Auto-Encoder and Generative Data Augmentation (CVPR 2025 · similarity 0.84)
- ConTex-Human: Free-View Rendering of Human from a Single Image with Texture-Consistent Synthesis (CVPR 2024 · similarity 0.84)
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.
