Research map / 3D vision / Depth and stereo
BridgeDepth: Bridging Monocular and Stereo Reasoning with Latent Alignment
Tongfan Guan, Jiaxin Guo, Chen Wang, Yun-Hui Liu
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 monocular-readout cross-attention (monocular priors into stereo aggregation) (ablation) | NMRF-s baseline · ETH3D zero-shot | BP-1: 3.7 -> 2.3 | Improved |
| add monocular-update (stereo evidence back into monocular stream) (ablation) | BridgeDepth · ETH3D zero-shot | BP-1: 2.3 -> 2.1 | Improved |
| add fine-grained 1/4 resolution alignment (ablation) | BridgeDepth · ETH3D zero-shot | BP-1: 2.1 -> 1.6 | Improved |
| add monocular loss supervision (ablation) | BridgeDepth · ETH3D zero-shot | BP-1: 1.6 -> 1.3 | Improved |
| add all monocular components (cost: slower inference) (ablation) | NMRF-s baseline · Inference time 960x540 | time (s): 0.044 -> 0.147 | Hurt |
| use iterative bidirectional alignment instead of post-hoc fusion (ablation) | BridgeDepth (DAv2-B) · ETH3D zero-shot | BP-1: 3.2 -> 1.4 | Improved |
| swap monocular backbone from DAv2-L to UniDepthV2 (ablation) | BridgeDepth · ETH3D zero-shot | BP-1: 1.3 -> 1.8 | Hurt |
| use BridgeDepth instead of NMRF in zero-shot evaluation | NMRF · Middlebury (Q) zero-shot | BP-2: 7.5 -> 4.3 | Improved |
| BridgeDepth versus FoundationStereo on KITTI-12 zero-shot | BridgeDepth · KITTI-12 zero-shot | D1: 3.6 vs 3.2 (FoundationStereo) | Hurt |
| enlarge stereo feature network (ConvNeXt-Tiny based BridgeDepth-L) | BridgeDepth · Scene Flow test | BP-1: 3.67 -> 3.25 | Improved |
Later papers that compared with this method
1 later accepted paper reports BridgeDepth as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- DispViT: Direct Stereo Disparity Regression with a Single-Stream Vision Transformer (ICLR 2026 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- MonSter — MonSter: Marry Monodepth to Stereo Unleashes Power (CVPR 2025) · better on 3 of 3 shared results
- Mocha-Stereo — MoCha-Stereo: Motif Channel Attention Network for Stereo Matching (CVPR 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Geometric Distillation from Rectified Stereo: Leveraging Epipolar Cues for Monocular Depth (ECCV 2026 · similarity 0.91)
- Single View Stereo Matching (CVPR 2018 · similarity 0.90)
- MonSter: Marry Monodepth to Stereo Unleashes Power (CVPR 2025 · similarity 0.90)
- MonoInstance: Enhancing Monocular Priors via Multi-view Instance Alignment for Neural Rendering and Reconstruction (CVPR 2025 · similarity 0.89)
- Stereo Anywhere: Robust Zero-Shot Deep Stereo Matching Even Where Either Stereo or Mono Fail (CVPR 2025 · similarity 0.89)
- RefineAny3D: Depth Refinement as Semantic Alignment for Monocular 3D Detection (NeurIPS 2026 · 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.
