AI Paper Atlas

Research map / 3D vision / 3D generation

Neural Scene De-Rendering

CVPR 2017 · 3D vision › 3D generation · approach: Semantic scene completion · 110 citations (OpenAlex)

Check your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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 segment proposals instead of box proposals (ablation)NSD · Abstract Scenereconstruction error (%): 28.12 -> 23.76Helped
Add segment proposals instead of box proposals (ablation)NSD · Minecraftreconstruction error (%): 11.42 -> 7.71Helped
Add segment proposals instead of box proposals (ablation)NSD · Abstract Sceneinference error (%): 47.85 -> 44.19Helped
Add segment proposals instead of box proposals (ablation)NSD · Minecraftinference error (%): 32.20 -> 32.11No clear effect
Add REINFORCE to segment-proposal model (ablation)NSD · Abstract Scenereconstruction error (%): 23.76 -> 22.44Helped
Add REINFORCE to segment-proposal model (ablation)NSD · Minecraftreconstruction error (%): 7.71 -> 5.73Helped
Add REINFORCE to segment-proposal model (ablation)NSD · Abstract Sceneinference error (%): 44.19 -> 45.09Hurt
Add REINFORCE to segment-proposal model (ablation)NSD · Minecraftinference error (%): 32.11 -> 28.79Helped
Add analysis-by-synthesis refinement to the REINFORCE model (ablation)NSD · Abstract Scenereconstruction error (%): 22.44 -> 21.55Helped
Add analysis-by-synthesis refinement to the REINFORCE model (ablation)NSD · Minecraftreconstruction error (%): 5.73 -> 5.05Helped
Use semi-supervised training with labels on 10% of training images, then unlabeled reconstruction fine-tuning (ablation)NSD · Abstract Scenereconstruction error (%): 21.55 vs 21.96No clear effect
Use semi-supervised training with labels on 10% of training images, then unlabeled reconstruction fine-tuning (ablation)NSD · Minecraftreconstruction error (%): 5.05 vs 7.62Hurt

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