Research map / 3D vision / NeRF and radiance fields
3D-Aware Conditional Image Synthesis
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 a 3D label-rendering branch (ablation) | pix2pix3D · Seg2Face | mIoU: N/A (0.43) -> 0.51 (0.52) | Helped |
| add a 3D label-rendering branch (ablation) | pix2pix3D · Edge2Car | AP: 0.45 (0.42) -> 0.63 (0.59) | Helped |
| add cross-view consistency loss (ablation) | pix2pix3D · Seg2Face | FVV Identity: 0.42 -> 0.36 | Helped |
| add cross-view consistency loss (ablation) | pix2pix3D · Edge2Car | FID: 9.42 -> 8.31 | Helped |
| add the image discriminator (ablation) | pix2pix3D · Seg2Face | FID: 15.32 -> 11.54 | Helped |
| add the conditional discriminator (ablation) | pix2pix3D · Seg2Face | mIoU: 0.37 (0.47) -> 0.51 (0.52) | Helped |
| concatenate RGB images into the conditional discriminator input (ablation) | pix2pix3D · Seg2Face | mIoU: 0.41 (0.40) -> 0.51 (0.52) | Helped |
| use pretrained EG3D weights to initialize parts of the model (ablation) | pix2pix3D · Seg2Face | FID: 11.54 -> 11.13 | Helped |
| sample random poses with probability 0 instead of 0.5 (ablation) | pix2pix3D · training pose sampling | image quality | Hurt |
| sample random poses with probability 1 instead of 0.5 (ablation) | pix2pix3D · training pose sampling | input alignment | Hurt |
| remove the image discriminator (ablation) | pix2pix3D · Seg2Face | FID: 11.54 -> 15.32 | Hurt |
| remove the conditional discriminator (ablation) | pix2pix3D · Seg2Face | mIoU: 0.51 (0.52) -> 0.37 (0.47) | Hurt |
Closest papers by meaning
- Towards Unsupervised Learning of Generative Models for 3D Controllable Image Synthesis (CVPR 2020 · similarity 0.89)
- Pix2NeRF: Unsupervised Conditional p-GAN for Single Image to Neural Radiance Fields Translation (CVPR 2022 · similarity 0.87)
- 3D-Aware Image Synthesis via Learning Structural and Textural Representations (CVPR 2022 · similarity 0.87)
- A Shading-Guided Generative Implicit Model for Shape-Accurate 3D-Aware Image Synthesis (NeurIPS 2021 · similarity 0.87)
- SIGNeRF: Scene Integrated Generation for Neural Radiance Fields (CVPR 2024 · similarity 0.85)
- Training-free Dense-Aligned Diffusion Guidance for Modular Conditional Image Synthesis (CVPR 2025 · similarity 0.85)
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.