Research map / 3D vision / Depth and stereo
Active Stereo Without Pattern Projector
Luca Bartolomei, Matteo Poggi, Fabio Tosi, Andrea Conti, Stefano Mattoccia
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 virtual pattern projection (random pointwise) to stereo input (ablation) | RAFT-Stereo · Midd-A (5% depth) | error rate >2px (%): 11.5 -> 5.2 | Improved |
| add virtual pattern projection (random pointwise) to stereo input (ablation) | rSGM · Midd-A (5% depth) | error rate >2px (%): 34.3 -> 20.6 | Improved |
| add alpha blending to pointwise patterns (ablation) | VPP with RAFT-Stereo · Midd-A (5% depth) | error rate >2px (%): 5.2 -> 5.8 | Hurt |
| project patterns on 3x3 patches with alpha blending (ablation) | VPP with rSGM · Midd-A (5% depth) | error rate >2px (%): 21.2 -> 16.7 | Improved |
| use FGD occlusion handling instead of BKGD projection (ablation) | VPP with rSGM · Midd-A (5% depth) | error rate >2px (%): 15.9 -> 15.3 | Improved |
| apply VPP to off-the-shelf RAFT-Stereo without retraining instead of guided stereo | RAFT-Stereo · Midd-14 (synthetic-trained) | error rate >2px (%): 11.58 -> 5.30 | Improved |
| train from scratch with guided stereo instead of VPP | RAFT-Stereo · Midd-14 (trained from scratch) | error rate >2px (%): 3.29 -> 4.14 | Mixed |
| combine guided stereo with VPP on RAFT-Stereo (ablation) | RAFT-Stereo-vpp · Midd-14 (synthetic-trained) | error rate >2px (%): 5.30 -> 6.52 | Hurt |
| combine guided stereo with VPP on rSGM (ablation) | rSGM-vpp · Midd-14 (synthetic-trained) | error rate >2px (%): 20.37 -> 19.55 | Improved |
| add VPP to LidarStereoNet and CCVNorm (ablation) | CCVNorm · Midd-14 (synthetic-trained) | error rate >2px (%): 12.49 -> 12.26 | Improved |
| apply VPP to off-the-shelf stereo networks with real LiDAR on KITTI | RAFT-Stereo · KITTI 142 (raw LiDAR) | error rate >2px (%): 8.74 -> 6.33 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- PSMNet — PSMNet: Position-Aware Stereo Merging Network for Room Layout Estimation (CVPR 2022) · 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
- Active Stereo in the Wild through Virtual Pattern Projection (IJCV 2025 · similarity 0.94)
- Active-Passive SimStereo - Benchmarking the Cross-Generalization Capabilities of Deep Learning-based Stereo Methods (NeurIPS 2022 · similarity 0.87)
- Polka Lines: Learning Structured Illumination and Reconstruction for Active Stereo (CVPR 2021 · similarity 0.87)
- Active Event-based Stereo Vision (CVPR 2025 · similarity 0.87)
- Toward Ultrafast Depth Sensing via Active Event-Based Stereo Vision (TPAMI 2026 · similarity 0.87)
- Stereo Anywhere: Robust Zero-Shot Deep Stereo Matching Even Where Either Stereo or Mono Fail (CVPR 2025 · similarity 0.83)
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.
