Research map / Detection and segmentation / 3D detection and driving
Wasserstein Distances for Stereo Disparity Estimation
Divyansh Garg, Yan Wang, Bharath Hariharan, Mark Campbell, Kilian Q. Weinberger, Wei-Lun Chao
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 continuous offsets and Wasserstein training to GANet Deep | GANet Deep · Scene Flow test | EPE: 0.78 -> 0.70 | Improved |
| Add continuous offsets and Wasserstein training to GANet Deep | GANet Deep · KITTI 2015 foreground | 3PE: 3.46 -> 3.20 | Improved |
| Add continuous offsets and Wasserstein training to GANet Deep | GANet Deep · KITTI 2015 all pixels | 3PE: 1.81 -> 1.92 | Mixed |
| Replace the depth estimator with CDN-SDN | DSGN · KITTI 3D test, car, moderate, IoU=0.7 | AP_3D: 52.2 -> 54.2 | Improved |
| Replace SDN with CDN-SDN | pseudo-LiDAR++ · KITTI 3D test, car, moderate, IoU=0.7 | AP_3D: 42.4 -> 44.9 | Improved |
| Add offsets while retaining mean regression and mean output (ablation) | PSMNet · Scene Flow test | EPE: 1.09 -> 1.04 | No clear effect |
| Use W1 loss and mode output without offsets (ablation) | PSMNet · Scene Flow test | EPE: 1.09 -> 1.20 | Hurt |
| Use W1 loss and mode output without offsets (ablation) | PSMNet · Scene Flow test | 1PE: 12.1 -> 10.5 | Improved |
| Combine offsets with W1 training and mode output (ablation) | PSMNet · Scene Flow test | EPE: 1.09 -> 0.98 | Improved |
| Replace W1 with KL divergence (ablation) | CDN-SDN · Scene Flow | RMSE: 1.81 -> 2.57 | Hurt |
| Replace W1 with W2^2 distance (ablation) | CDN-SDN · Scene Flow | RMSE: 1.81 -> 1.91 | Hurt |
| Set disparity bin size to 1 instead of 2 (ablation) | CDN-PSMNet · Scene Flow test | EPE: 0.98 -> 1.22 | Hurt |
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.
- DSGN — DSGN: Deep Stereo Geometry Network for 3D Object Detection (CVPR 2020) · better on 2 of 2 shared results
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
- Triangulation Learning Network: From Monocular to Stereo 3D Object Detection (CVPR 2019 · similarity 0.88)
- Disp R-CNN: Stereo 3D Object Detection via Shape Prior Guided Instance Disparity Estimation (CVPR 2020 · similarity 0.87)
- A Depth Complementary Framework for Monocular 3D Object Detection in Autonomous Driving (TPAMI 2026 · similarity 0.87)
- Shape Prior Guided Instance Disparity Estimation for 3D Object Detection (TPAMI 2021 · similarity 0.87)
- IDA-3D: Instance-Depth-Aware 3D Object Detection From Stereo Vision for Autonomous Driving (CVPR 2020 · similarity 0.87)
- Stereographic Spherical Sliced Wasserstein Distances (ICML 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.
