Inverting The Generator Of A Generative Adversarial Network

Abstract

Generative adversarial networks (GANs) learn to synthesise new samples from a high-dimensional distribution by passing samples drawn from a latent space through a generative network. When the high-dimensional distribution describes images of a particular data set, the network should learn to generate visually similar image samples for latent variables that are close to each other in the latent space. For tasks such as image retrieval and image classification, it may be useful to exploit the arrangement of the latent space by projecting images into it, and using this as a representation for discriminative tasks. GANs often consist of multiple layers of non-linear computations, making them very difficult to invert. This paper introduces techniques for projecting image samples into the latent space using any pre-trained GAN, provided that the computational graph is available. We evaluate these techniques on both MNIST digits and Omniglot handwritten characters. In the case of MNIST digits

Stats

• citations222
• S2 citations—
• github stars0
• HF likes0
• heat score17.61
• arxiv keycreswell2016inverting

Related papers

• Drop The GAN: In Defense Of Patches Nearest Neighbors As Single Image Generative Models (2021)12.87
• Adversarial Training For Sketch Retrieval (2016)11.19
• Generative Adversarial Nets For Information Retrieval: Fundamentals And Advances (2018)0.00
• Dark Side Augmentation: Generating Diverse Night Examples For Metric Learning (2023)8.34
• Semi-supervised Learning With Bidirectional Gans (2018)4.52
• Binary Generative Adversarial Networks For Image Retrieval (2017)18.34
• Regularizing Deep Hashing Networks Using GAN Generated Fake Images (2018)0.00
• Unsupervised Adversarial Attacks On Deep Feature-based Retrieval With GAN (2019)0.00