DenseFace：基于密度感知概率匹配的人脸识别偏见缓解
原标题：DenseFace: Bias Mitigation in Face Recognition via Density-Aware Probabilistic Matching
AI 摘要
该论文提出 DenseFace，一种基于 von Mises-Fisher 分布密度感知的概率匹配方法，用于在无需重新训练的前提下减少预训练人脸识别模型中的种族偏见。方法通过建模人脸嵌入的局部密度并调整相似度分数，在多种网络架构、训练数据集和损失函数上均能持续降低偏见，同时保持识别准确率。作者还采用 NIST 评估协议，以固定阈值下的误匹配率作为主要偏见指标，并指出 RFW 协议在跨种族匹配场景中的不足。
正文节选
DenseFace: Bias Mitigation in Face Recognition via Density-Aware Probabilistic Matching Abstract Despite steady progress in face recognition, current face recognition models still suffer from significant demographic biases. While approaches for bias mitigation have been proposed, existing methods often impose constraints on the training procedure and result in the degradation of recognition accuracy. To address this issue, we here introduce a method that reduces racial bias in pre-trained face