Research map / 3D vision / Depth and stereo
PTC-Depth: Pose-Refined Monocular Depth Estimation with Temporal Consistency
Leezy Han, Seunggyu Kim, Dongseok Shim, Hyeonbeom Lee
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 Bayesian temporal fusion (ablation) | PTC-Depth · TartanAir (ablation sequence) | delta<1.25: 0.517 -> 0.714 | Improved |
| add Bayesian temporal fusion (ablation) | PTC-Depth · Roadside (ablation sequence) | TAE: 12.99 -> 6.27 | Improved |
| add Bayesian temporal fusion (ablation) | PTC-Depth · Roadside (ablation sequence) | AbsRel: 0.428 -> 0.265 | Improved |
| add Bayesian temporal fusion (ablation) | PTC-Depth · KITTI (ablation sequence) | TAE: 6.28 -> 4.79 | Improved |
| add segment-wise scale consolidation (ablation) | PTC-Depth · TartanAir (ablation sequence) | AbsRel: 0.274 -> 0.218 | Improved |
| add segment-wise scale consolidation (ablation) | PTC-Depth · KITTI (ablation sequence) | TAE: 4.38 -> 4.79 | Hurt |
| add segment-wise scale consolidation (ablation) | PTC-Depth · Roadside (ablation sequence) | delta<1.25: 0.746 -> 0.743 | No clear effect |
| use ground-truth rotation instead of estimated pose (ablation) | PTC-Depth · MS2 thermal (odometry ablation) | AbsRel: 0.513 -> 0.254 | Improved |
| use full ground-truth pose instead of estimated pose (ablation) | PTC-Depth · TartanAir (odometry ablation) | AbsRel: 0.218 -> 0.172 | Improved |
| use full ground-truth pose instead of estimated pose (ablation) | PTC-Depth · MS2 thermal (odometry ablation) | AbsRel: 0.513 -> 0.974 | Hurt |
| use PTC-Depth instead of UniDepth on out-of-distribution data | UniDepth · Roadside (RGB), full range 0-80 m | AbsRel: 0.465 -> 0.309 | Improved |
| use PTC-Depth instead of UniDepth on in-distribution data | UniDepth · KITTI (RGB), full range 0-80 m | AbsRel: 0.047 -> 0.137 | Hurt |
Closest papers by meaning
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.89)
- RePoseD: Efficient Relative Pose Estimation With Known Depth Information (ICCV 2025 · similarity 0.89)
- Relative Pose Estimation through Affine Corrections of Monocular Depth Priors (CVPR 2025 · similarity 0.87)
- Adaptive Fusion of Single-View and Multi-View Depth for Autonomous Driving (CVPR 2024 · similarity 0.87)
- Seurat: From Moving Points to Depth (CVPR 2025 · similarity 0.86)
- A Simple yet Universal Framework for Depth Completion (NeurIPS 2024 · similarity 0.86)
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.
