从频率域解析人脸识别偏差，发现不同种族依赖不同频段特征。
Frequency Matters: Explaining Biases of Face Recognition in the Frequency Domain
- 用频域分析法揭示人脸识别模型对不同频段的依赖差异。
- 非裔样本更依赖低频信息，亚裔样本则偏高频特征。
- 为理解算法偏见提供新视角，适合关注公平性研究者。
人脸识别模型在不同人口群体间存在性能差异，其原因因深度学习模型结构复杂而难以明确。已有研究指出性别与种族偏差可能源于发型、妆容或面部毛发等语义因素。受近期关于卷积神经网络中频率模式重要性的启发，本文采用先进的频域解释方法分析人脸识别中的偏差。大量实验表明，不同种族样本的人脸识别模型所依赖的频率成分存在显著差异，其中非裔样本更依赖低频信息，而亚裔样本则更依赖高频特征。该发现为理解模型偏见提供了新的视觉化解释路径。
原文摘要 · Abstract (English)
Face recognition (FR) models are vulnerable to performance variations across demographic groups. The causes for these performance differences are unclear due to the highly complex deep learning-based structure of face recognition models. Several works aimed at exploring possible roots of gender and ethnicity bias, identifying semantic reasons such as hairstyle, make-up, or facial hair as possible sources. Motivated by recent discoveries of the importance of frequency patterns in convolutional neural networks, we explain bias in face recognition using state-of-the-art frequency-based explanations. Our extensive results show that different frequencies are important to FR models depending on the ethnicity of the samples.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。