Research map / 3D vision / Depth and stereo
Digging Into Self-Supervised Monocular Depth Estimation
Clement Godard, Oisin Mac Aodha, Michael Firman, Gabriel J. Brostow
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 minimum reprojection loss (ablation) | Monodepth2 · KITTI Eigen split, monocular, pretrained | Abs Rel: 0.117 -> 0.115 | Improved |
| add auto-masking (ablation) | Monodepth2 · KITTI Eigen split, monocular, pretrained | Abs Rel: 0.120 -> 0.115 | Improved |
| add full-resolution multi-scale sampling (ablation) | Monodepth2 · KITTI Eigen split, monocular, pretrained | Abs Rel: 0.117 -> 0.115 | Improved |
| add minimum reprojection loss to baseline (ablation) | Baseline · KITTI Eigen split, monocular, pretrained | Abs Rel: 0.140 -> 0.122 | Improved |
| add auto-masking to baseline (ablation) | Baseline · KITTI Eigen split, monocular, pretrained | Abs Rel: 0.140 -> 0.124 | Improved |
| add full-res multi-scale to baseline (ablation) | Baseline · KITTI Eigen split, monocular, pretrained | Abs Rel: 0.140 -> 0.124 | Improved |
| replace auto-masking with predictive mask (ablation) | Monodepth2 · KITTI Eigen split, monocular, pretrained | Abs Rel: 0.115 -> 0.123 | Hurt |
| add SSIM to photometric loss (ablation) | Monodepth2 · KITTI Eigen split, monocular, pretrained | Abs Rel: 0.118 -> 0.115 | Improved |
| add ImageNet pretraining (ablation) | Monodepth2 · KITTI Eigen split, monocular | Abs Rel: 0.132 -> 0.115 | Improved |
| add all three contributions without pretraining (ablation) | Baseline w/o pt · KITTI Eigen split, monocular, from scratch | Abs Rel: 0.150 -> 0.132 | Improved |
| train on full Eigen split with Monodepth2 components (ablation) | Baseline · KITTI full Eigen split, monocular | Abs Rel: 0.146 -> 0.116 | Improved |
| increase input resolution from 416x128 to 640x192 (ablation) | Monodepth2 · KITTI Eigen split, monocular, pretrained | Abs Rel: 0.128 -> 0.115 | Improved |
Closest papers by meaning
- On the Uncertainty of Self-Supervised Monocular Depth Estimation (CVPR 2020 · similarity 0.94)
- Self-Supervised Monocular Trained Depth Estimation Using Self-Attention and Discrete Disparity Volume (CVPR 2020 · similarity 0.94)
- Self-Supervised Monocular Depth Hints (ICCV 2019 · similarity 0.93)
- Excavating the Potential Capacity of Self-Supervised Monocular Depth Estimation (ICCV 2021 · similarity 0.93)
- The Temporal Opportunist: Self-Supervised Multi-Frame Monocular Depth (CVPR 2021 · similarity 0.92)
- SC-DepthV3: Robust Self-Supervised Monocular Depth Estimation for Dynamic Scenes (TPAMI 2023 · similarity 0.92)
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.
