VIP内容

题目: Diverse Image Generation via Self-Conditioned GANs

摘要:

本文介绍了一个简单但有效的无监督方法,以产生现实和多样化的图像,并且训练了一个类条件GAN模型,而不使用手动注释的类标签。相反,模型的条件是标签自动聚类在鉴别器的特征空间。集群步骤自动发现不同的模式,并显式地要求生成器覆盖它们。在标准模式基准测试上的实验表明,该方法在寻址模式崩溃时优于其他几种竞争的方法。并且该方法在ImageNet和Places365这样的大规模数据集上也有很好的表现,与以前的方法相比,提高了图像多样性和标准质量指标。

成为VIP会员查看完整内容
0
26

最新内容

Various controls over the generated data can be extracted from the latent space of a pre-trained GAN, as it implicitly encodes the semantics of the training data. The discovered controls allow to vary semantic attributes in the generated images but usually lead to entangled edits that affect multiple attributes at the same time. Supervised approaches typically sample and annotate a collection of latent codes, then train classifiers in the latent space to identify the controls. Since the data generated by GANs reflects the biases of the original dataset, so do the resulting semantic controls. We propose to address disentanglement by subsampling the generated data to remove over-represented co-occuring attributes thus balancing the semantics of the dataset before training the classifiers. We demonstrate the effectiveness of this approach by extracting disentangled linear directions for face manipulation on two popular GAN architectures, PGGAN and StyleGAN, and two datasets, CelebAHQ and FFHQ. We show that this approach outperforms state-of-the-art classifier-based methods while avoiding the need for disentanglement-enforcing post-processing.

0
0
下载
预览

最新论文

Various controls over the generated data can be extracted from the latent space of a pre-trained GAN, as it implicitly encodes the semantics of the training data. The discovered controls allow to vary semantic attributes in the generated images but usually lead to entangled edits that affect multiple attributes at the same time. Supervised approaches typically sample and annotate a collection of latent codes, then train classifiers in the latent space to identify the controls. Since the data generated by GANs reflects the biases of the original dataset, so do the resulting semantic controls. We propose to address disentanglement by subsampling the generated data to remove over-represented co-occuring attributes thus balancing the semantics of the dataset before training the classifiers. We demonstrate the effectiveness of this approach by extracting disentangled linear directions for face manipulation on two popular GAN architectures, PGGAN and StyleGAN, and two datasets, CelebAHQ and FFHQ. We show that this approach outperforms state-of-the-art classifier-based methods while avoiding the need for disentanglement-enforcing post-processing.

0
0
下载
预览
参考链接
Top
微信扫码咨询专知VIP会员