Research map / 3D vision / Depth and stereo
Every Pixel Counts ++: Joint Learning of Geometry and Motion with 3D Holistic Understanding
Chenxu Luo, Zhenheng Yang, Peng Wang, Yang Wang, Wei Xu, Ram Nevatia, Alan Yuille
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 depth consistency fine-tuning (ablation) | EPC++ mono · KITTI Eigen depth | Abs Rel: 0.151 -> 0.146 | Improved |
| add depth consistency fine-tuning (Sq Rel) (ablation) | EPC++ mono · KITTI Eigen depth | Sq Rel: 1.448 -> 1.065 | Improved |
| add flow consistency without visibility mask (ablation) | EPC++ mono · KITTI Eigen depth | Abs Rel: 0.151 -> 0.148 | No clear effect |
| add visibility mask to flow consistency (ablation) | EPC++ mono · KITTI Eigen depth | Abs Rel: 0.148 -> 0.144 | Improved |
| joint training of depth and flow networks (ablation) | EPC++ mono · KITTI Eigen depth | Abs Rel: 0.146 -> 0.156 | Hurt |
| two iterations of adaptive depth-flow consistency training (ablation) | EPC++ mono · KITTI Eigen depth | Abs Rel: 0.144 -> 0.141 | Improved |
| add stereo training samples | EPC++ · KITTI Eigen depth | Abs Rel: 0.141 -> 0.127 | Improved |
| add depth consistency in stereo setting (ablation) | EPC++ stereo · KITTI Eigen depth | Abs Rel: 0.141 -> 0.134 | Improved |
| add depth-flow consistency, masks and adaptive training (ablation) | EPC++ mono · MPI-Sintel depth | Abs Rel: 0.866 -> 0.524 | Improved |
| joint fine-tune flow with depth rather than flow only (ablation) | OptFlowNet · KITTI 2015 training flow | EPE noc: 3.66 -> 4.00 | Hurt |
| finetune flow on all regions with depth-derived flow (ablation) | OptFlowNet · KITTI 2015 training flow | EPE occ: 23.07 -> 16.20 | Improved |
| finetune only on non-visible regions rather than visible-rigid regions (ablation) | OptFlowNet · KITTI 2015 training flow | EPE all: 7.17 -> 5.84 | Improved |
Closest papers by meaning
- Feature-Level Collaboration: Joint Unsupervised Learning of Optical Flow, Stereo Depth and Camera Motion (CVPR 2021 · similarity 0.91)
- GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Pose (CVPR 2018 · similarity 0.90)
- Deep Scene Flow Learning: From 2D Images to 3D Point Clouds (TPAMI 2023 · similarity 0.90)
- Seurat: From Moving Points to Depth (CVPR 2025 · similarity 0.88)
- SpatialTracker: Tracking Any 2D Pixels in 3D Space (CVPR 2024 · similarity 0.88)
- MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motion (ICLR 2025 · similarity 0.88)
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.
