AI Paper Atlas

Research map / 3D vision / Human pose estimation

HAP: Structure-Aware Masked Image Modeling for Human-Centric Perception

Junkun Yuan, Xinyu Zhang, Hao Zhou, Jian Wang, Zhongwei Qiu, Zhiyin Shao, Shaofeng Zhang, Sifan Long, Kun Kuang, Kun Yao, Junyu Han, Errui Ding and 3 more

NeurIPS 2023 · 3D vision › Human pose estimation · approach: Heatmap and coordinate pose encoding · 5 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
Use human structure settings: high scaling ratio, intermediate masking ratio, and block-wise masking (ablation)MAE pre-training on LUPerson · MSMT17mAP: 68.9 -> 74.1Improved
Add human-part-guided mask sampling (ablation)HAP-0 · MSMT17mAP: 74.1 -> 75.1Improved
Add structure-invariant alignment loss (ablation)HAP-0 · MSMT17mAP: 74.1 -> 75.6Improved
Use attention-map guidance instead of part guidance (ablation)HAP mask sampling · MSMT17mAP: 75.6 -> 71.7Hurt
Guide masking with keypoints rather than whole human parts (ablation)HAP mask sampling · MSMT17mAP: 75.6 -> 75.9No clear effect
Select a fixed 3 parts instead of sampling the number of parts uniformly from 0 to 6 (ablation)HAP mask sampling · MSMT17mAP: 76.4 -> 75.4Hurt
Select a fixed 6 parts instead of sampling the number of parts uniformly from 0 to 6 (ablation)HAP mask sampling · MSMT17mAP: 76.4 -> 74.4Hurt
Predict keypoints instead of reconstructing image pixels (ablation)HAP decoder objective · MSMT17mAP: 50.3 vs. 72.2Hurt
Use OpenPose instead of ViTPose to extract keypoints (ablation)HAP pre-training · MSMT17mAP: 72.0 vs. 72.2No clear effect
Set alignment-loss weight γ to 0 (ablation)HAP · MSMT17mAP: 76.4 -> 75.1Hurt
Align a masked view with a global unmasked view instead of two masked views (ablation)HAP structure-invariant alignment · MSMT17mAP: 76.4 -> 74.5Hurt
Use SimSiam-style alignment without negative samples (ablation)HAP alignment loss · MSMT17mAP: 76.4 vs. 76.1No 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