AI Paper Atlas

Research map / 3D vision / Human pose estimation

An Empirical Study of the Collapsing Problem in Semi-Supervised 2D Human Pose Estimation

Rongchang Xie, Chunyu Wang, Wenjun Zeng, Yizhou Wang

ICCV 2021 · 3D vision › Human pose estimation · approach: Heatmap and coordinate pose encoding · 38 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 easy-hard augmentation pair with gradient only through the hard branch instead of plain consistency (ablation)consistency-based SSL · COCO val (1K labels)AP: 0.9 -> 41.5Improved
detach the teacher (easy) signal so gradients flow only through the student (ablation)Ours · COCOcollapseImproved
use easy augmentation for both branches or hard for the teacher instead of easy-hard (ablation)Ours · COCOheatmap responseHurt
learn dual networks instead of a single network (ablation)Ours · COCO val (1K labels)AP: 38.5 -> 41.5Improved
add Joint Cutout as hard augmentation to dual model (ablation)Ours (Dual) · COCO val (1K labels)AP: 41.5 -> 44.6Improved
add Joint Cutout to single model (ablation)Ours (Single) · COCO val (1K labels)AP: 38.5 -> 42.1Improved
use Joint Cutout instead of RandAug as hard augmentation (ablation)Ours (Dual) · COCO val (1K labels)AP: 43.7 -> 44.6Improved
apply hard augmentation in supervised training only (ablation)Supervised ResNet18 · COCO valAP: 51.1 -> 52.1Mixed
pair ResNet18 with HRNet w48 partner instead of another ResNet18 (ablation)Ours (Dual) · COCO val (1K labels)AP: 41.5 -> 48.7Improved
add unlabeled COCO WILD images with many labelsOurs · COCO val, ResNet50AP: 70.9 -> 73.9Improved
add unlabeled COCO WILD images on strong HRNetW48Ours · COCO val, HRNetW48AP: 77.2 -> 79.2Improved
add semi-supervised learning to DARK on test-devOurs (+DARK) · COCO test-devAP: 76.2 -> 77.2Improved

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