arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2002.12093cs.CV
收藏

减少人脸识别的地理性能差异

Reducing Geographic Performance Differential for Face Recognition

Onfido, UK(Onfido公司,英国)

Martins Bruveris, Jochem Gietema, Pouria Mortazavian, Mohan Mahadevan

摘要:

随着人脸识别算法变得越来越准确,并得到更广泛的部署,确保算法对每个人都同样有效变得越来越重要。我们研究了地理性能差异-不同国家的错误接受率和错误拒绝率的差异-当比较自拍照与身份证件照片时。我们展示了如何使用采样策略来减轻地理性能差异,尽管数据集存在很大的不平衡。使用普通领域自适应策略来微调特定于域的文档自拍数据上的人脸识别CNN,提高了模型在此类数据上的性能,但是,在存在不平衡的训练数据的情况下,也显著增加了人口统计偏差。然后,我们展示了如何通过采用采样策略来平衡训练过程来减轻这种影响。

英文摘要:

As face recognition algorithms become more accurate and get deployed more widely, it becomes increasingly important to ensure that the algorithms work equally well for everyone. We study the geographic performance differentials-differences in false acceptance and false rejection rates across different countries-when comparing selfies against photos from ID documents. We show how to mitigate geographic performance differentials using sampling strategies despite large imbalances in the dataset. Using vanilla domain adaptation strategies to fine-tune a face recognition CNN on domain-specific doc-selfie data improves the performance of the model on such data, but, in the presence of imbalanced training data, also significantly increases the demographic bias. We then show how to mitigate this effect by employing sampling strategies to balance the training procedure.

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。

↑