AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

End-to-End One-Shot Human Parsing

Haoyu He, Jing Zhang, Bohan Zhuang, Jianfei Cai, Dacheng Tao

TPAMI 2023 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Fine-tuning strategies for few-shot

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 NPM to AGM with equal fixed weight (DML without weight shifting) (ablation)AGM-only EOP-Net · ATR-OS Fold 1C_novel mIoU: 0.4 -> 13.1Improved
add weight-shifting strategy to DML (ablation)DML w/o WS · ATR-OS Fold 1C_novel mIoU: 13.1 -> 21.6Improved
add weight-shifting strategy to DML (ablation)DML w/o WS · ATR-OS Fold 1C_human mIoU: 33.3 -> 58.6Improved
add coarse-grained metric space (end-to-end framework) (ablation)DML · ATR-OS Fold 1C_novel mIoU: 21.6 -> 30.2Improved
add prototype-level contrastive loss (ablation)DML + CG & FG · ATR-OS Fold 1C_novel mIoU: 30.2 -> 31.1Improved
use end-to-end instead of progressive framework (ablation)POPNet-style framework · ATR-OS Fold 1C_human mIoU: 55.3 -> 61.1Improved
use NPM only instead of AGM only (ablation)EOP-Net · ATR-OS Fold 1C_human mIoU: 51.6 -> 16.5Hurt
raise momentum coefficient alpha from 0.8 to 0.999 (ablation)momentum-updated prototypes · ATR-OS Fold 1C_novel mIoU: 25.7 -> 31.1Improved
raise momentum coefficient alpha from 0.999 to 0.9999 (ablation)momentum-updated prototypes · ATR-OS Fold 1C_novel mIoU: 31.1 -> 31.0No clear effect
use 62-layer instead of 74-layer embedding network (ablation)EOP-Net · ATR-OS Fold 1C_novel mIoU: 30.4 -> 31.1Improved
replace progressive POPNet with end-to-end EOP-NetPOPNet · CIHP-OS k-wayC_novel mIoU: 17.7 -> 22.8Improved
replace POPNet with EOP-Net in 1-way settingPOPNet · CIHP-OS 1-wayC_novel mIoU: 35.7 -> 30.9Hurt

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