Fairer Analysis and Demographically Balanced Face Generation for Fairer Face Verification
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915369831104512 |
|---|---|
| author | Fournier-Montgieux, Alexandre Soumm, Michael Popescu, Adrian Luvison, Bertrand Borgne, Hervé Le |
| author_facet | Fournier-Montgieux, Alexandre Soumm, Michael Popescu, Adrian Luvison, Bertrand Borgne, Hervé Le |
| contents | Face recognition and verification are two computer vision tasks whose performances have advanced with the introduction of deep representations. However, ethical, legal, and technical challenges due to the sensitive nature of face data and biases in real-world training datasets hinder their development. Generative AI addresses privacy by creating fictitious identities, but fairness problems remain. Using the existing DCFace SOTA framework, we introduce a new controlled generation pipeline that improves fairness. Through classical fairness metrics and a proposed in-depth statistical analysis based on logit models and ANOVA, we show that our generation pipeline improves fairness more than other bias mitigation approaches while slightly improving raw performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_03349 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Fairer Analysis and Demographically Balanced Face Generation for Fairer Face Verification Fournier-Montgieux, Alexandre Soumm, Michael Popescu, Adrian Luvison, Bertrand Borgne, Hervé Le Computer Vision and Pattern Recognition Face recognition and verification are two computer vision tasks whose performances have advanced with the introduction of deep representations. However, ethical, legal, and technical challenges due to the sensitive nature of face data and biases in real-world training datasets hinder their development. Generative AI addresses privacy by creating fictitious identities, but fairness problems remain. Using the existing DCFace SOTA framework, we introduce a new controlled generation pipeline that improves fairness. Through classical fairness metrics and a proposed in-depth statistical analysis based on logit models and ANOVA, we show that our generation pipeline improves fairness more than other bias mitigation approaches while slightly improving raw performance. |
| title | Fairer Analysis and Demographically Balanced Face Generation for Fairer Face Verification |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2412.03349 |