Research map / 3D vision / NeRF and radiance fields
3D-Aware Indoor Scene Synthesis with Depth Priors
Zifan Shi, Yujun Shen, Jiapeng Zhu, Dit-Yan Yeung, Qifeng Chen
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 depth prediction losses (real and fake) (ablation) | DepthGAN · LSUN bedroom 128x128 | FID (Depth): 26.518 -> 18.874 | Improved |
| add depth prediction losses (real and fake) (ablation) | DepthGAN · LSUN bedroom 128x128 | Rotation consistency (RC): 0.711 -> 0.530 | Improved |
| add depth prediction loss on fake images (ablation) | DepthGAN · LSUN bedroom 128x128 | FID: 5.441 -> 4.040 | Improved |
| add RGB rotation consistency loss (ablation) | DepthGAN · LSUN bedroom 128x128 | Rotation consistency (RC): 0.716 -> 0.530 | Improved |
| add RGB rotation consistency loss (ablation) | DepthGAN · LSUN bedroom 128x128 | FID: 5.038 -> 4.040 | Improved |
| add depth rotation consistency loss (ablation) | DepthGAN · LSUN bedroom 128x128 | Rotation precision (RP): 0.152 -> 0.040 | Improved |
| add depth rotation consistency loss (ablation) | DepthGAN · LSUN bedroom 128x128 | FID (Depth): 8.315 -> 18.874 | Hurt |
| add depth-feature conditioning of the appearance renderer (ablation) | DepthGAN · LSUN bedroom 128x128 | FID: 24.062 -> 4.040 | Improved |
| add depth-feature conditioning of the appearance renderer (ablation) | DepthGAN · LSUN bedroom 128x128 | FID (Depth): 119.917 -> 18.874 | Improved |
| add conditioning and rotation losses (ablation) | DepthGAN · LSUN bedroom 128x128 | FID: 2.793 -> 4.040 | Hurt |
| use depth prior with dual-path generator instead of GIRAFFE | DepthGAN · LSUN bedroom 256x256 | FID: 44.232 -> 4.797 | Improved |
| give baselines depth supervision (GRAF to GRAF-RGBD) | GRAF · LSUN bedroom 256x256 | FID: 66.856 -> 194.260 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- GIRAFFE — GIRAFFE: Representing Scenes As Compositional Generative Neural Feature Fields (CVPR 2021) · better on 3 of 3 shared results
- pi-GAN — Pi-GAN: Periodic Implicit Generative Adversarial Networks for 3D-Aware Image Synthesis (CVPR 2021) · better on 3 of 3 shared results
- GRAF — GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis (NeurIPS 2020) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Indoor Scene Generation From a Collection of Semantic-Segmented Depth Images (ICCV 2021 · similarity 0.92)
- Simple and Effective Synthesis of Indoor 3D Scenes (AAAI 2023 · similarity 0.91)
- Fast and Flexible Indoor Scene Synthesis via Deep Convolutional Generative Models (CVPR 2019 · similarity 0.90)
- Conditional 360-degree Image Synthesis for Immersive Indoor Scene Decoration (ICCV 2023 · similarity 0.90)
- CAD: Photorealistic 3D Generation via Adversarial Distillation (CVPR 2024 · similarity 0.87)
- InstructScene: Instruction-Driven 3D Indoor Scene Synthesis with Semantic Graph Prior (ICLR 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.
