AI Paper Atlas

Research map / Detection and segmentation / CNN features

Learning Robust Facial Landmark Detection via Hierarchical Structured Ensemble

Xu Zou, Sheng Zhong, Luxin Yan, Xiangyun Zhao, Jiahuan Zhou, Ying Wu

ICCV 2019 · Detection and segmentation › CNN features · approach: Local features and context in CNNs · 65 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 HSLE structural constraintsHG baseline · 300W Fullmean iod-norm error: 3.77 -> 3.28Improved
add HSLE structural constraintsDU-Net baseline · 300W Fullmean iod-norm error: 3.47 -> 3.30Improved
add HSLE structural constraintsMerget et al. baseline · 300W Fullmean iod-norm error: 4.26 -> 3.70Improved
add HSLE structural constraintsMerget et al. baseline · 300W Challengemean iod-norm error: 5.69 -> 5.69No clear effect
add HSLE structural constraintsHG baseline · 300W Fullmean ipd-norm error: 5.27 -> 4.59Improved
add HSLE structural constraintsHG baseline · 300W Commonmedian iod-norm error: 3.17 -> 2.72Improved
add HSLE structural constraints4-stack hourglass · AFLW-Fullmean iod-norm error: 7.84 -> 7.52Improved
add HSLE structural constraints8-stack hourglass · AFLW-Fullmean iod-norm error: 7.64 -> 7.24Improved
replace manual bi-directional tree structure with learned HSLE (ablation)4-stack hourglass · AFLW-Fullmean iod-norm error: 7.64 -> 7.52Improved
replace manual bi-directional tree structure with learned HSLE (ablation)8-stack hourglass · AFLW-Fullmean iod-norm error: 7.56 -> 7.24Improved
increase number of ensembles from 4 to 8 (ablation)HSLE-S4 · AFLW-Fullmean iod-norm error: 7.60 -> 7.47Improved
increase number of ensembles from 8 to 14 (ablation)HSLE-S4 · AFLW-Fullmean iod-norm error: 7.47 -> 7.52No clear effect

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