AI Paper Atlas

Research map / Image generation / Editing and style transfer

SynFace: Face Recognition With Synthetic Data

Haibo Qiu, Baosheng Yu, Dihong Gong, Zhifeng Li, Wei Liu, Dacheng Tao

ICCV 2021 · Image generation › Editing and style transfer · approach: Diffusion face generation and editing · 160 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 identity mixup during training (ablation)SynFace Syn_10K_50 · LFWaccuracy (%): 88.98 -> 91.97Improved
add identity mixup during training (ablation)SynFace 1K_50 · LFWaccuracy (%): 83.85 -> 87.53Improved
add identity mixup during training (ablation)SynFace 10K_5 · LFWaccuracy (%): 88.22 -> 88.32No clear effect
increase mixup level from none to phi 0.6-1.0 (intra-class variation) (ablation)SynFace 5K_50 · LFWaccuracy (%): 88.75 -> 90.95Improved
balance class distribution (2K_UB1 to 2K_50) (ablation)SynFace · LFW long-tailedaccuracy (%): 83.80 -> 86.18Improved
increase identities (width) from 1K to 5K (ablation)SynFace · LFWaccuracy (%): 83.85 -> 88.75Improved
increase identities (width) from 5K to 10K (ablation)SynFace · LFWaccuracy (%): 88.75 -> 88.98No clear effect
increase samples per identity (depth) from 2 to 20 (ablation)SynFace 10K · LFWaccuracy (%): 78.85 -> 89.90Improved
increase samples per identity (depth) from 20 to 50 (ablation)SynFace 10K · LFWaccuracy (%): 89.90 -> 88.98No clear effect
vary pose and illumination rather than fixing attributes (ablation)SynFace 5K_50 · LFWaccuracy (%): 74.55 -> 73.72Improved
add 1K real identities with 10 samples each via domain mixupSynFace Syn_10K_50 with IM · LFWaccuracy (%): 91.97 -> 92.28Improved
add domain mixup to real data (2K ids, 10 samples)Real_2K_10 · LFWaccuracy (%): 91.22 -> 95.78Improved

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