AI Paper Atlas

Research map / 3D vision / 3D generation

IKEA-Manual: Seeing Shape Assembly Step by Step

Ruocheng Wang, Yunzhi Zhang, Jiayuan Mao, Ran Zhang, Chin-Yi Cheng, Jiajun Wu

NeurIPS 2022 · 3D vision › 3D generation · approach: Scene-aware human motion synthesis · 7 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
Recursively group parts by geometric similarity rather than connect all parts in one stepAssembly plan generation · IKEA-ManualHard matching F1: 10.78 -> 16.30Improved
Recursively group parts by geometric similarity rather than connect all parts in one stepAssembly plan generation · IKEA-ManualSimple matching precision: 100 -> 44.90Hurt
Use geometric similarity alone to build assembly-plan groupsGeoCluster · IKEA-ManualHard matching F1: 16.30Hurt
Fine-tune the pretrained segmentation model on IKEA-Manual examplesPointNet-conditioned U-Net · IKEA-Manual part-conditioned segmentation testIoU: 25.31Mixed
Optimize pose from masks using differentiable rendering instead of learned pose predictionPart-conditioned pose estimation · IKEA-ManualRotation error: 82.79 -> 64.81Improved
Optimize pose from masks using differentiable rendering instead of learned pose predictionPart-conditioned pose estimation · IKEA-ManualCA: 49.53 -> 51.4Improved
Use differentiable-rendering optimization at inferenceSoftRas pose estimation · IKEA-ManualInference time (s): <10^{-5} -> 243Hurt
Evaluate a PartNet-trained assembly model on realistic IKEA-Manual parts rather than PartNet partsPretrained part-assembly models · Chair part assemblyPart Accuracy: RGL-Net*: 49.06 -> 3.99; Huang et al.: 39.96 -> 6.90Hurt
Evaluate a PartNet-trained assembly model on realistic IKEA-Manual parts rather than PartNet partsPretrained part-assembly models · Chair part assemblyShape CD: RGL-Net*: 0.0087 -> 0.0508; Huang et al.: 0.0091 -> 0.0151Hurt

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