arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2304.07060cs.CV
收藏

DCFace:基于双条件扩散模型的合成人脸生成

DCFace: Synthetic Face Generation with Dual Condition Diffusion Model

Michigan State University(密歇根州立大学)

Minchul Kim, Feng Liu, Anil Jain, Xiaoming Liu

AI总结:

本文提出基于扩散模型的双条件人脸生成器DCFace,通过逐块风格提取器和依赖时间步的ID损失结合主体外观与风格条件,实现同一主体在不同风格下的精确人脸合成,并在多个基准上显著提升人脸识别准确率。

摘要:

为训练人脸识别模型生成合成数据集具有挑战性,因为数据集的生成不仅涉及创建高保真图像,还涉及在遵循真实图像条件分布的不同因素(例如姿态、光照、表情、年龄和遮挡的变化)下生成同一主体的多张图像。先前的工作已经研究了使用GAN或3D模型生成合成数据集。在这项工作中,我们从结合主体外观(ID)和外部因素(风格)条件的角度来处理该问题。这两个条件提供了一种直接控制类间和类内变化的方法。为此,我们提出了一种基于扩散模型的双条件人脸生成器(DCFace)。我们新颖的逐块风格提取器和依赖时间步的ID损失使DCFace能够在精确控制下持续生成同一主体在不同风格下的人脸图像。在5个测试数据集LFW、CFP-FP、CPLFW、AgeDB和CALFW中的4个上,使用所提出的DCFace生成的合成图像训练的人脸识别模型与先前工作相比,平均验证准确率提高了6.11%。代码可在 https://github.com/mk-minchul/dcface 获取。

英文摘要:

Generating synthetic datasets for training face recognition models is challenging because dataset generation entails more than creating high fidelity images. It involves generating multiple images of same subjects under different factors (\textit{e.g.}, variations in pose, illumination, expression, aging and occlusion) which follows the real image conditional distribution. Previous works have studied the generation of synthetic datasets using GAN or 3D models. In this work, we approach the problem from the aspect of combining subject appearance (ID) and external factor (style) conditions. These two conditions provide a direct way to control the inter-class and intra-class variations. To this end, we propose a Dual Condition Face Generator (DCFace) based on a diffusion model. Our novel Patch-wise style extractor and Time-step dependent ID loss enables DCFace to consistently produce face images of the same subject under different styles with precise control. Face recognition models trained on synthetic images from the proposed DCFace provide higher verification accuracies compared to previous works by $6.11\%$ on average in $4$ out of $5$ test datasets, LFW, CFP-FP, CPLFW, AgeDB and CALFW. Code is available at https://github.com/mk-minchul/dcface

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。

↑