AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

HAL3D: Hierarchical Active Learning for Fine-Grained 3D Part Labeling

Fenggen Yu, Yiming Qian, Francisca Gil-Ureta, Brian Jackson, Eric Bennett, Hao Zhang

ICCV 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 3 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 hierarchical labeling to active learning (ablation)HAL3D · PartNet Chairlabeling time (h): 5.99 -> 4.34Improved
add hierarchical labeling to active learning (ablation)HAL3D · PartNet Chairpart label accuracy: 89.84 -> 94.13Improved
add symmetry checking (ablation)HAL3D · PartNet Chairlabeling time (h): 5.21 -> 4.34Improved
add active learning fine-tuning to hierarchical symmetry-aware labeling (ablation)HAL3D · PartNet Chairlabeling time (h): 6.37 -> 4.34Improved
add proposal network before manual modification (ablation)manual labeling · PartNet Chairlabeling time (h): 22.05 -> 8.65Improved
use proposal network instead of NGSP for initial labelsHAL3D · PartNet inference time, Chairseconds per shape: 5.410 -> 0.012Improved
use proposal network instead of NGSP for initial labelsHAL3D · PartNet Chairpart label accuracy: 68.98 -> 67.45Hurt
use HAL3D instead of correcting PartNet or NGSP predictions manuallyABO annotation · ABO Chairlabeling time (h): 28.30 -> 6.46Improved

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