AI Paper Atlas

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

Demodalizing Face Recognition with Synthetic Samples

Zhonghua Zhai, Pengju Yang, Xiaofeng Zhang, Maji Huang, Haijing Cheng, Xuejun Yan, Chunmao Wang, Shiliang Pu

AAAI 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Fair and quality-aware face recognition · 7 citations (OpenAlex)

Read the paper

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 synthetic samples with explicit identity/non-identity disentanglingCosFace · IJB-ATAR@FAR=1e-4: 0.61718 -> 0.86662Improved
Use synthetic samples with explicit identity/non-identity disentanglingCosFace · IJB-CTAR@FAR=1e-5: 0.55689 -> 0.84819Improved
Use synthetic samples with explicit identity/non-identity disentanglingCosFace · CQ-IJB-CTAR@FAR=1e-5: 0.26264 -> 0.59166Improved
Use synthetic samples with explicit identity/non-identity disentanglingCosFace · CP-IJB-CTAR@FAR=1e-5: 0.10385 -> 0.5393Improved
Use synthetic samples with implicit meta-learning demodalizationCosFace · IJB-ATAR@FAR=1e-3: 0.85133 -> 0.91923Improved
Use synthetic samples with implicit meta-learning demodalizationCosFace · CP-IJB-CTAR@FAR=1e-5: 0.10385 -> 0.157767Improved
Use synthetic samples with feature-filter demodalizationCosFace · CQ-IJB-CTAR@FAR=1e-5: 0.26264 -> 0.37303Improved
Use synthetic samples with feature-filter demodalizationCosFace · CP-IJB-CTAR@FAR=1e-5: 0.10385 -> 0.17101Improved
Train with synthetic samples using DMFR rather than without synthetic samplesCosFace · LFWaccuracy: 0.9975 -> 0.9972No clear effect
Train with synthetic samples using DMFR rather than without synthetic samplesCosFace · MegaFaceaccuracy: 0.7893 -> 0.7660Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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