arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2204.14079cs.CVcs.AIcs.LG
收藏

修复噪声:解纠缠源特征用于StyleGAN的迁移学习

Fix the Noise: Disentangling Source Feature for Transfer Learning of StyleGAN

KAIST(韩国科学技术院)

Dongyeun Lee, Jae Young Lee, Doyeon Kim, Jaehyun Choi, Junmo Kim

摘要:

StyleGAN的迁移学习最近显示出解决各种任务的巨大潜力,特别是在领域翻译中。以前的方法在迁移学习过程中通过交换或冻结权重来利用源模型,然而,它们在视觉质量和控制源特征方面具有局限性。换句话说,它们需要额外的模型,这些模型在计算上要求很高,并且具有限制的控制步骤,从而阻止了平滑过渡。在本文中,我们提出了一种新的方法来克服这些限制。而不是交换或冻结,我们引入了一个简单的特征匹配损失,以提高生成质量。此外,为了控制源特征的程度,我们用所提出的策略FixNoise训练目标模型,以仅在目标特征空间的解纠缠子空间中保留源特征。由于特征空间的解纠缠,我们的方法可以平滑地控制在一个单一的模型中的源特征的程度。大量的实验表明,该方法可以生成更一致和真实的图像比以前的作品。

英文摘要:

Transfer learning of StyleGAN has recently shown great potential to solve diverse tasks, especially in domain translation. Previous methods utilized a source model by swapping or freezing weights during transfer learning, however, they have limitations on visual quality and controlling source features. In other words, they require additional models that are computationally demanding and have restricted control steps that prevent a smooth transition. In this paper, we propose a new approach to overcome these limitations. Instead of swapping or freezing, we introduce a simple feature matching loss to improve generation quality. In addition, to control the degree of source features, we train a target model with the proposed strategy, FixNoise, to preserve the source features only in a disentangled subspace of a target feature space. Owing to the disentangled feature space, our method can smoothly control the degree of the source features in a single model. Extensive experiments demonstrate that the proposed method can generate more consistent and realistic images than previous works.

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。

↑