AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Disentangled Representation for Age-Invariant Face Recognition: A Mutual Information Minimization Perspective

Xuege Hou, Yali Li, Shengjin Wang

ICCV 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Fair and quality-aware face recognition · 34 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 mutual information minimization between identity and age features (ablation)MT-MIM (ArcFace baseline) · MORPH Album 2Rank-1 accuracy (%): 99.00 -> 99.43Improved
add mutual information minimization (ablation)MT-MIM · CACD (2004-2006)Rank-1 accuracy (%): 91.81 -> 92.63Improved
add mutual information minimization (ablation)MT-MIM · CACD (2007-2009)Rank-1 accuracy (%): 93.27 -> 93.95Improved
add mutual information minimization (ablation)MT-MIM · CACD (2010-2012)Rank-1 accuracy (%): 95.35 -> 96.09Improved
add mutual information minimization (ablation)MT-MIM · FG-NET (leave-one-out)Rank-1 accuracy (%): 93.20 -> 94.21Improved
add mutual information minimization (ablation)MT-MIM · FG-NET (MF2)Rank-1 accuracy (%): 60.26 -> 61.12Improved
add mutual information minimization (ablation)MT-MIM · AgeDB-30Rank-1 accuracy (%): 95.11 -> 96.10Improved
add age classification supervision alone (without MI minimization) (ablation)ArcFace baseline · CACD (2007-2009)Rank-1 accuracy (%): 93.27 -> 93.15No clear effect
add age classification supervision alone (ablation)ArcFace baseline · FG-NET (MF2)Rank-1 accuracy (%): 60.26 -> 60.04No clear effect
add age classification supervision alone (ablation)ArcFace baseline · AgeDB-30Rank-1 accuracy (%): 95.11 -> 95.40Improved
add MI minimization on top of age supervision (ablation)MT-MIM · MORPH Album 2Rank-1 accuracy (%): 99.10 -> 99.43Improved
fine-tune on CACD after general face pretrainingCNN baseline · CACD (2004-2006)MAP (%): 68.21 -> 91.81Improved

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