arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2108.07960cs.CV
收藏

SynFace:基于合成数据的人脸识别

SynFace: Face Recognition with Synthetic Data

JD Explore Academy(京东探索研究院);The University of Sydney(悉尼大学);Tencent Data Platform(腾讯数据平台)

Haibo Qiu, Baosheng Yu, Dihong Gong, Zhifeng Li, Wei Liu, Dacheng Tao

摘要:

随着最近深度神经网络的成功,人脸识别取得了显着进展。然而,收集大规模的真实世界的人脸识别训练数据已被证明是具有挑战性的,特别是由于标签噪声和隐私问题。同时,现有的人脸识别数据集通常是从网络图像中收集的,缺乏对属性的详细注释(例如,姿态和表情),因此不同属性对人脸识别的影响研究得很少。在本文中,我们解决了上述问题,在人脸识别使用合成人脸图像,即,SynFace具体来说,我们首先探讨了最近使用合成和真实人脸图像训练的最先进的人脸识别模型之间的性能差距。然后,我们分析了性能差距背后的根本原因,例如,差的类内变化和合成和真实人脸图像之间的域间隙。受此启发,我们设计了具有身份混淆(IM)和域混淆(DM)的SynFace,以缓解上述性能差距,展示了合成数据在人脸识别中的巨大潜力。此外,通过可控的人脸合成模型,我们可以很容易地管理合成人脸生成的不同因素,包括姿态,表情,光照,身份的数量和每个身份的样本。因此,我们也对合成人脸图像进行了系统的实证分析,为如何有效地利用合成数据进行人脸识别提供了一些见解。

英文摘要:

With the recent success of deep neural networks, remarkable progress has been achieved on face recognition. However, collecting large-scale real-world training data for face recognition has turned out to be challenging, especially due to the label noise and privacy issues. Meanwhile, existing face recognition datasets are usually collected from web images, lacking detailed annotations on attributes (e.g., pose and expression), so the influences of different attributes on face recognition have been poorly investigated. In this paper, we address the above-mentioned issues in face recognition using synthetic face images, i.e., SynFace. Specifically, we first explore the performance gap between recent state-of-the-art face recognition models trained with synthetic and real face images. We then analyze the underlying causes behind the performance gap, e.g., the poor intra-class variations and the domain gap between synthetic and real face images. Inspired by this, we devise the SynFace with identity mixup (IM) and domain mixup (DM) to mitigate the above performance gap, demonstrating the great potentials of synthetic data for face recognition. Furthermore, with the controllable face synthesis model, we can easily manage different factors of synthetic face generation, including pose, expression, illumination, the number of identities, and samples per identity. Therefore, we also perform a systematically empirical analysis on synthetic face images to provide some insights on how to effectively utilize synthetic data for face recognition.

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。

↑