发表机构
VisionLabs; Amazon Web Services; Mohamed bin Zayed University of Artificial Intelligence（VisionLabs; 亚马逊云服务; 穆罕默德·本·扎耶德人工智能大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对人脸识别中的种族偏差问题，提出基于冯·米塞斯-费舍尔分布密度感知的概率匹配方法DenseFace，在不降低精度且无需重训练的情况下有效缓解偏差。
AI 中文摘要
尽管人脸识别取得了稳步进展，但当前的人脸识别模型仍然存在显著的群体统计偏差。虽然已有多种偏差缓解方法被提出，但现有方法往往对训练过程施加约束，并导致识别精度的下降。为解决这一问题，我们在此提出一种方法，在不牺牲精度的情况下减少预训练人脸识别模型中的种族偏差。为此，我们使用冯·米塞斯-费舍尔（MF）分布对每个人的面部嵌入进行建模。接下来，我们观察人口统计属性与MF分布密度之间的依赖关系，并提出DenseFace，一种考虑MF分布差异的概率人脸匹配过程。我们的大量实验表明，DenseFace能够持续减少在多种网络架构、训练数据集和损失函数上表现强劲的人脸识别模型中的种族偏差。值得注意的是，DenseFace保持了识别精度，且无需对底层人脸识别模型进行重新训练。我们的工作还考察了先前采用的偏差度量，并提出了建议。
英文摘要
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 recognition models without compromising their accuracy. To this end, we model face embeddings of each person by von Mises-Fisher (MF) distribution. We next observe the dependency between demographic attributes and the density of MF distributions, and propose DenseFace, a probabilistic face matching procedure that accounts for differences in MF distributions. Our extensive experiments demonstrate DenseFace to consistently reduce racial bias in strong face recognition models varying in network architectures, training datasets and loss functions. Notably, DenseFace preserves recognition accuracy and requires no retraining of the underlying face recognition model. Our work also investigates previously adopted bias measures and makes suggestions.
Comments13 pages, 10 figures. Accepted at IEEE/IAPR International Joint Conference on Biometrics (IJCB) 2026
