AI Paper Atlas

Research map / 3D vision / 3D generation

SSLayout360: Semi-Supervised Indoor Layout Estimation From 360deg Panorama

Phi Vu Tran

CVPR 2021 · 3D vision › 3D generation · approach: Generative indoor scene synthesis

Check 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 unlabeled-data consistency trainingHorizonNet · PanoContext, 20 labels3D IoU (%): 61.48 ± 2.07 -> 63.05 ± 0.65Improved
Add unlabeled-data consistency trainingHorizonNet · PanoContext, 963 labels3D IoU (%): 83.55 ± 0.31 -> 83.30 ± 0.53Hurt
Add unlabeled-data consistency trainingHorizonNet · Stanford 2D-3D, 20 labels3D IoU (%): 62.20 ± 3.98 -> 71.60 ± 2.04Improved
Add unlabeled-data consistency trainingHorizonNet · Stanford 2D-3D, 916 labels3D IoU (%): 82.79 ± 0.90 -> 84.66 ± 0.57Improved
Add unlabeled-data consistency trainingHorizonNet · MatterportLayout (4–18 corners), 50 labels3D IoU (%): 63.44 ± 0.56 -> 67.42 ± 0.24Improved
Add unlabeled-data consistency trainingHorizonNet · MatterportLayout (4–18 corners), 1,650 labels3D IoU (%): 79.12 ± 0.37 -> 80.33 ± 0.48Improved
Use EMA teacher parameters instead of shared student-teacher parameters (ablation)SSLayout360 · PanoContext validation set3D IoUImproved
Set consistency loss weight to λ = 1 (ablation)SSLayout360 · PanoContext validation set3D IoUImproved
Use a T@30% consistency-loss ramp-up (ablation)SSLayout360 · PanoContext validation set3D IoUImproved
Use EMA decay α = 0.999 (ablation)SSLayout360 · PanoContext validation set3D IoUImproved
Use squared L2 loss instead of cross-entropy for wall-wall corner predictions (ablation)SSLayout360 · Layout estimation experimentsaccuracyImproved

Later papers that compared with this method

1 later accepted paper reports SSLayout360* as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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