Research map / 3D vision / Depth and stereo
Two-in-One Depth: Bridging the Gap Between Monocular and Binocular Self-Supervised Depth Estimation
Zhengming Zhou, Qiulei Dong
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 Monocular Feature Matching module instead of concatenation module (ablation) | TiO-Depth · KITTI 2015 binocular | D1: 15.952 -> 6.684 | Improved |
| use MFM instead of cross-attention without SE layer (ablation) | TiO-Depth · KITTI 2015 binocular | D1: 7.421 -> 6.684 | Improved |
| increase number of MFM modules from 1 to 3 (ablation) | TiO-Depth · KITTI 2015 binocular | D1: 8.784 -> 6.684 | Improved |
| add disparity guidance loss, cost volume loss and occlusion mask (ablation) | TiO-Depth stereo loss · KITTI 2015 binocular | D1: 7.159 -> 6.684 | Improved |
| add final branches in SDFA modules (ablation) | TiO-Depth monocular · KITTI Eigen test | Abs. Rel.: 0.098 -> 0.085 | Improved |
| train jointly on two tasks without distillation (steps 1+2) (ablation) | TiO-Depth monocular · KITTI Eigen test | Abs. Rel.: 0.088 -> 0.088 | No clear effect |
| add step 3 distillation with hybrid probability volume (ablation) | TiO-Depth monocular · KITTI Eigen test | Abs. Rel.: 0.088 -> 0.085 | Improved |
| use hybrid volume P_h instead of stereo volume P_s as teacher (ablation) | TiO-Depth monocular · KITTI Eigen test | Abs. Rel.: 0.086 -> 0.085 | Improved |
| binocular input instead of monocular input in cross-dataset test | TiO-Depth · Cityscapes (trained on KITTI) | Abs. Rel.: 0.120 -> 0.066 | Improved |
Closest papers by meaning
- Self-Supervised Monocular Trained Depth Estimation Using Self-Attention and Discrete Disparity Volume (CVPR 2020 · similarity 0.91)
- BRNet: Exploring Comprehensive Features for Monocular Depth Estimation (ECCV 2022 · similarity 0.91)
- Self-Supervised Monocular Depth Estimation by Direction-aware Cumulative Convolution Network (ICCV 2023 · similarity 0.91)
- On the Synergies between Machine Learning and Binocular Stereo for Depth Estimation from Images: a Survey (TPAMI 2021 · similarity 0.91)
- Self-Supervised Monocular Depth Hints (ICCV 2019 · similarity 0.90)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.89)
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.
