AI Paper Atlas

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

NeurIPS 2020 · Detection and segmentation › 3D detection and driving · approach: Monocular 3D object detection · 47 citations (OpenAlex)

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 changedWhereMeasuredEffect
Add continuous offsets and Wasserstein training to GANet DeepGANet Deep · Scene Flow testEPE: 0.78 -> 0.70Improved
Add continuous offsets and Wasserstein training to GANet DeepGANet Deep · KITTI 2015 foreground3PE: 3.46 -> 3.20Improved
Add continuous offsets and Wasserstein training to GANet DeepGANet Deep · KITTI 2015 all pixels3PE: 1.81 -> 1.92Mixed
Replace the depth estimator with CDN-SDNDSGN · KITTI 3D test, car, moderate, IoU=0.7AP_3D: 52.2 -> 54.2Improved
Replace SDN with CDN-SDNpseudo-LiDAR++ · KITTI 3D test, car, moderate, IoU=0.7AP_3D: 42.4 -> 44.9Improved
Add offsets while retaining mean regression and mean output (ablation)PSMNet · Scene Flow testEPE: 1.09 -> 1.04No clear effect
Use W1 loss and mode output without offsets (ablation)PSMNet · Scene Flow testEPE: 1.09 -> 1.20Hurt
Use W1 loss and mode output without offsets (ablation)PSMNet · Scene Flow test1PE: 12.1 -> 10.5Improved
Combine offsets with W1 training and mode output (ablation)PSMNet · Scene Flow testEPE: 1.09 -> 0.98Improved
Replace W1 with KL divergence (ablation)CDN-SDN · Scene FlowRMSE: 1.81 -> 2.57Hurt
Replace W1 with W2^2 distance (ablation)CDN-SDN · Scene FlowRMSE: 1.81 -> 1.91Hurt
Set disparity bin size to 1 instead of 2 (ablation)CDN-PSMNet · Scene Flow testEPE: 0.98 -> 1.22Hurt

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.

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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy