Research map / 3D vision / Depth and stereo
HRDFuse: Monocular 360deg Depth Estimation by Collaboratively Learning Holistic-With-Regional Depth Distributions
Hao Ai, Zidong Cao, Yan-Pei Cao, Ying Shan, Lin Wang
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 HRDFuse with 256×256 TP patches instead of OmniFusion (2-iter) | 360° monocular depth estimation · 3D60 | Abs Rel: 0.0430 -> 0.0358 | Improved |
| Use HRDFuse with 256×256 TP patches instead of OmniFusion (2-iter) | 360° monocular depth estimation · Stanford2D3D | Abs Rel: 0.0950 -> 0.0935 | Improved |
| Use HRDFuse with 128×128 TP patches instead of PanoFormer | 360° monocular depth estimation · Matterport3D | Abs Rel: 0.0904 -> 0.0967 | Hurt |
| Combine ERP and TP branches, then add SFA instead of geometric fusion (ablation) | HRDFuse · Matterport3D | FPS: 2.82 -> 6.21 | Improved |
| Add CDDC to the ERP+TP model with SFA (ablation) | HRDFuse · Matterport3D | Abs Rel: 0.0991 -> 0.0967 | Improved |
| Use only the ERP branch rather than combining ERP and TP (ablation) | HRDFuse · Matterport3D | Abs Rel: 0.0967 -> 0.1028 | Hurt |
| Use 46 rather than 18 TP patches with 128×128 patches (ablation) | HRDFuse · Matterport3D | Abs Rel: 0.0967 -> 0.1232 | Hurt |
| Use 46 rather than 18 TP patches with 256×256 patches (ablation) | HRDFuse · Matterport3D | Abs Rel: 0.0981 -> 0.0966 | Improved |
| Reduce TP patch FoV from 80° to 60° (ablation) | HRDFuse with 128×128 TP patches · Matterport3D | Abs Rel: 0.0967 -> 0.0986 | Hurt |
| Increase TP patch FoV from 80° to 100° (ablation) | HRDFuse with 128×128 TP patches · Matterport3D | Abs Rel: 0.0967 -> 0.0970 | Hurt |
| Use 100 rather than 20 depth histogram bins (ablation) | HRDFuse · Matterport3D | Abs Rel: 0.0997 -> 0.0967 | Improved |
| Increase depth histogram bins from 100 to 150 (ablation) | HRDFuse · Matterport3D | Abs Rel: 0.0967 -> 0.0997 | Hurt |
Later papers that compared with this method
3 later accepted papers report HRDFuse as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- DepthMaster: Unified Monocular Depth Estimation for Perspective and Panoramic Images (NeurIPS 2026 · better on 2 of 2 shared results)
- One2Scene: Geometric Consistent Explorable 3D Scene Generation from a Single Image (ICLR 2026 · better on 1 of 1 shared result)
- Depth Anywhere: Enhancing 360 Monocular Depth Estimation via Perspective Distillation and Unlabeled Data Augmentation (NeurIPS 2024 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- PanoFormer — PanoFormer: Panorama Transformer for Indoor 360° Depth Estimation (ECCV 2022) · better on 0 of 1 shared result
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
- HR-Depth: High Resolution Self-Supervised Monocular Depth Estimation (AAAI 2021 · similarity 0.90)
- 360MonoDepth: High-Resolution 360deg Monocular Depth Estimation (CVPR 2022 · similarity 0.89)
- Learning Depth from Single Monocular Images Using Deep Convolutional Neural Fields (TPAMI 2016 · similarity 0.88)
- Depth from a Single Image by Harmonizing Overcomplete Local Network Predictions (NeurIPS 2016 · similarity 0.88)
- Self-Supervised Monocular Depth Estimation by Direction-aware Cumulative Convolution Network (ICCV 2023 · similarity 0.88)
- DME: Unveiling the Bias for Better Generalized Monocular Depth Estimation (AAAI 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.
