Research map / 3D vision / Depth and stereo
Unsupervised Monocular Depth Estimation With Left-Right Consistency
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 left-right disparity consistency (ablation) | Monocular depth estimation on KITTI · KITTI Eigen split | Abs Rel: 0.152 -> 0.148 | Helped |
| Add left-right disparity consistency (ablation) | Monocular depth estimation · KITTI 2015 stereo 200 training set disparity images | D1-all: 30.318 -> 30.272 | Helped |
| Replace Deep3D image formation with the proposed image formation model (ablation) | Monocular depth estimation · KITTI 2015 stereo 200 training set disparity images | D1-all: 66.85 -> 30.272 | Helped |
| Add smoothness constraint to Deep3D image formation (ablation) | Monocular depth estimation · KITTI 2015 stereo 200 training set disparity images | D1-all: 66.85 -> 59.64 | Helped |
| Pre-train on Cityscapes and train/fine-tune on KITTI (ablation) | Monocular depth estimation · KITTI Eigen split | Abs Rel: 0.148 -> 0.124 | Helped |
| Train on Cityscapes alone rather than KITTI (ablation) | Monocular depth estimation · KITTI 2015 stereo 200 training set disparity images | Abs Rel: 0.124 -> 0.699 | Hurt |
| Add post-processing using flipped-image predictions (ablation) | Monocular depth estimation · KITTI Eigen split | Abs Rel: 0.124 -> 0.118 | Helped |
| Replace the encoder with ResNet50 (ablation) | Post-processed monocular depth model · KITTI Eigen split | Abs Rel: 0.118 -> 0.114 | Helped |
| Use both stereo views as network input instead of one (ablation) | Depth estimation · KITTI 2015 stereo 200 training set disparity images | D1-all: 30.272 -> 9.194 | Helped |
| Add post-processing using flipped-image predictions (ablation) | Monocular depth estimation · Make3D test set | Abs Rel: 0.535 -> 0.443 | Helped |
| Use batch normalization (ablation) | Monocular depth estimation network · KITTI | accuracy | No clear effect |
| Optimize all four output scales simultaneously rather than progressively (ablation) | Monocular depth estimation network · Training | convergence stability | Helped |
Closest papers by meaning
- Single View Stereo Matching (CVPR 2018 · similarity 0.92)
- Learning Monocular Depth Estimation Infusing Traditional Stereo Knowledge (CVPR 2019 · similarity 0.91)
- Bilateral Cyclic Constraint and Adaptive Regularization for Unsupervised Monocular Depth Prediction (CVPR 2019 · similarity 0.91)
- Refine and Distill: Exploiting Cycle-Inconsistency and Knowledge Distillation for Unsupervised Monocular Depth Estimation (CVPR 2019 · similarity 0.90)
- Single Image Depth Estimation Trained via Depth From Defocus Cues (CVPR 2019 · similarity 0.90)
- Self-Supervised Monocular Trained Depth Estimation Using Self-Attention and Discrete Disparity Volume (CVPR 2020 · similarity 0.90)
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.