Research map / 3D vision / Depth and stereo
Geometric Structure Based and Regularized Depth Estimation From 360 Indoor Imagery
Lei Jin, Yanyu Xu, Jia Zheng, Junfei Zhang, Rui Tang, Shugong Xu, Jingyi Yu, Shenghua Gao
Check 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 corner, boundary, and plane structure as both prior and regularizer (ablation) | Ours-baseline depth estimation · Stanford 2D-3D-S | RMS: 0.472 -> 0.421 | Improved |
| add corner information (ablation) | Geometric structure based depth estimation · Stanford 2D-3D-S | RMS: 0.425 -> 0.421 | Improved |
| add boundary information (ablation) | Geometric structure based depth estimation · Stanford 2D-3D-S | RMS: 0.436 -> 0.421 | Improved |
| add plane information (ablation) | Geometric structure based depth estimation · Stanford 2D-3D-S | RMS: 0.429 -> 0.421 | Improved |
| add structure as a prior (ablation) | Depth estimation with geometric structure regularizer · Stanford 2D-3D-S | RMS: 0.448 -> 0.421 | Improved |
| add structure as a regularizer (ablation) | Depth estimation with geometric structure prior · Stanford 2D-3D-S | RMS: 0.446 -> 0.421 | Improved |
| add geometric structure information (ablation) | Depth estimation on synthetic rooms without furniture · Synthetic dataset, without furniture | RMS: 0.668 -> 0.642 | Improved |
| add geometric structure information (ablation) | Depth estimation on synthetic rooms with furniture · Synthetic dataset, with furniture | RMS: 0.721 -> 0.666 | Improved |
| add the attention module (ablation) | Depth estimation trained with furniture and tested without furniture · Synthetic dataset, without furniture | RMS: 0.784 -> 0.730 | Improved |
| add geometric structure prior, furniture attention, and structure regularization | Depth estimation on Stanford 2D-3D-S · Stanford 2D-3D-S | RMS[m]: 0.472 -> 0.421 | Improved |
| use the proposed geometric representation for counterfactual depth | Counterfactual depth estimation · Counterfactual depth | RMS: 0.845 -> 0.823 | Improved |
Closest papers by meaning
- Toward Practical Monocular Indoor Depth Estimation (CVPR 2022 · similarity 0.89)
- GeoLayout: Geometry Driven Room Layout Estimation Based on Depth Maps of Planes (ECCV 2020 · similarity 0.89)
- Joint Graph-Based Depth Refinement and Normal Estimation (CVPR 2020 · similarity 0.88)
- Neural 3D Scene Reconstruction With Indoor Planar Priors (TPAMI 2024 · similarity 0.88)
- Neural 3D Scene Reconstruction With the Manhattan-World Assumption (CVPR 2022 · similarity 0.88)
- StructDepth: Leveraging the Structural Regularities for Self-Supervised Indoor Depth Estimation (ICCV 2021 · similarity 0.88)
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.
