Research map / 3D vision / Depth and stereo
Depth Estimation From Indoor Panoramas With Neural Scene Representation
Wenjie Chang, Yueyi Zhang, Zhiwei Xiong
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 |
|---|---|---|---|
| use Sphere Embedding instead of Cartesian [x, y, z] coordinates (ablation) | the color network · Matterport3D | MRE: 0.3606 -> 0.0258 | Improved |
| use Sphere Embedding instead of scaled [x/s, y/s, z/s] coordinates (ablation) | the color network · Matterport3D | MRE: 0.4228 -> 0.0258 | Improved |
| use Sphere Embedding instead of embedded [θ, ϕ, 1/ρ] coordinates (ablation) | the color network · Matterport3D | MRE: 0.1073 -> 0.0258 | Improved |
| add the proposed Manhattan World geometry-network initialization (ablation) | the proposed method · Matterport3D | MRE: 0.1615 -> 0.0258 | Improved |
| use the proposed geometry-network initialization instead of Sphere initialization (ablation) | the geometry network · Matterport3D | MRE: 0.0430 -> 0.0258 | Improved |
| add geometric consistency loss based on surface normals (ablation) | the proposed method · Matterport3D | MSE: 0.0592 -> 0.0532 | Improved |
| use 3 panorama views instead of 2 | the proposed method · Matterport3D | MRE: 0.0713 -> 0.0277 | Improved |
| use Ours (3-view) instead of 360SD-Net | Matterport3D depth estimation · Matterport3D | MSE: 0.0740 -> 0.0525 | Improved |
| use Ours (2-view) instead of ADfSM | Matterport3D depth estimation · Matterport3D | MSE: 0.1985 -> 0.1944 | Improved |
Closest papers by meaning
- PanoFormer: Panorama Transformer for Indoor 360° Depth Estimation (ECCV 2022 · similarity 0.90)
- SliceNet: Deep Dense Depth Estimation From a Single Indoor Panorama Using a Slice-Based Representation (CVPR 2021 · similarity 0.89)
- Joint 3D Layout and Depth Prediction from a Single Indoor Panorama Image (ECCV 2020 · similarity 0.88)
- Geometric Exploitation for Indoor Panoramic Semantic Segmentation (NeurIPS 2024 · similarity 0.85)
- NeRFPrior: Learning Neural Radiance Field as a Prior for Indoor Scene Reconstruction (CVPR 2025 · similarity 0.85)
- NC-SDF: Enhancing Indoor Scene Reconstruction Using Neural SDFs with View-Dependent Normal Compensation (CVPR 2024 · 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.
