面向更公平人脸验证的更公平分析与人口统计学平衡人脸生成
Fairer Analysis and Demographically Balanced Face Generation for Fairer Face Verification
AI总结:
本文基于 DCFace 框架提出一种受控人脸生成流程,结合 logit 模型与 ANOVA 统计分析,在略微提升原始性能的同时实现比现有偏差缓解方法更优的人脸验证公平性。
摘要:
人脸识别和人脸验证是两项计算机视觉任务,其性能随着深度表示的引入而提升。然而,由于人脸数据的敏感性以及现实世界训练数据集中的偏差,伦理、法律和技术方面的挑战阻碍了它们的发展。生成式人工智能通过创建虚构身份来解决隐私问题,但公平性问题仍然存在。我们使用现有的 DCFace SOTA 框架,引入了一种新的受控生成流程来改善公平性。通过经典公平性指标以及基于 logit 模型和 ANOVA 的深入统计分析方法,我们表明,与其他偏差缓解方法相比,我们的生成流程在略微提升原始性能的同时,更大程度地改善了公平性。
英文摘要:
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.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。