Research map / 3D vision / Depth and stereo
ACDNet: Adaptively Combined Dilated Convolution for Monocular Panorama Depth Estimation
Chuanqing Zhuang, Zhengda Lu, Yiqun Wang, Jun Xiao, Ying 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 ACDNet instead of the strongest reported baseline, SliceNet | Monocular panorama depth estimation · Stanford2D3D | RMSE: 0.3509 -> 0.3410 | Improved |
| Use ACDNet instead of UniFuse | Monocular panorama depth estimation · Matterport3D | MAE: 0.2814 -> 0.2670 | Improved |
| Use ACDNet instead of HoHoNet | Monocular panorama depth estimation · Structured3D | MAE: 0.0549 -> 0.0454 | Improved |
| Add ACDConv with adaptive channel-wise fusion rather than regular convolution (ablation) | ACDNet · Stanford2D3D | MAE: 0.2104 -> 0.1870 | Improved |
| Add adaptive channel-wise fusion rather than simple averaging (ablation) | ACDConv · Stanford2D3D | MAE: 0.2037 -> 0.1870 | Improved |
| Use channel-wise fusion rather than row-wise or pixel-wise fusion (ablation) | ACDConv · Stanford2D3D | MAE: 0.2096 -> 0.1870 | Improved |
| Use ACDConv rather than ASPP (ablation) | ACDNet · Stanford2D3D | MAE: 0.1990 -> 0.1870 | Improved |
| Combine dilation along both axes rather than only one axis (ablation) | ACDConv · Stanford2D3D | MAE: 0.1953 -> 0.1870; 0.1965 -> 0.1870 | Improved |
| Use four dilation settings instead of two or three (ablation) | ACDConv · Stanford2D3D | MAE: 0.2038 -> 0.1870; 0.1971 -> 0.1870 | Improved |
| Increase the number of dilation settings from four to five (ablation) | ACDConv · Stanford2D3D | MAE: 0.1870 -> 0.1963 | Hurt |
| Use circular padding rather than zero padding or left-right padding (ablation) | ACDNet · Stanford2D3D | MAE: 0.1948 -> 0.1870; 0.1935 -> 0.1870 | Improved |
| Use a larger ResNet backbone rather than ResNet18 (ablation) | ACDNet · Stanford2D3D | MAE: 0.2309 -> 0.2044 -> 0.1870 | Improved |
Later papers that compared with this method
1 later accepted paper reports ACDNet as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Distortion and Uncertainty Aware Loss for Panoramic Depth Completion (ICML 2023 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SliceNet — SliceNet: Deep Dense Depth Estimation From a Single Indoor Panorama Using a Slice-Based Representation (CVPR 2021) · better on 2 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
- Distortion-Aware Convolutional Filters for Dense Prediction in Panoramic Images (ECCV 2018 · similarity 0.87)
- Patch-Wise Attention Network for Monocular Depth Estimation (AAAI 2021 · similarity 0.87)
- Monocular Depth Estimation Using Relative Depth Maps (CVPR 2019 · similarity 0.87)
- SliceNet: Deep Dense Depth Estimation From a Single Indoor Panorama Using a Slice-Based Representation (CVPR 2021 · similarity 0.87)
- DepthMaster: Unified Monocular Depth Estimation for Perspective and Panoramic Images (NeurIPS 2026 · similarity 0.86)
- DCDepth: Progressive Monocular Depth Estimation in Discrete Cosine Domain (NeurIPS 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.
