On Leveraging Pretrained GANs for Generation with Limited Data

TL;DR

This paper explores leveraging pretrained GANs and adaptive filter modulation to improve image generation quality in scenarios with limited training data, demonstrating effective transfer learning techniques.

Contribution

It introduces a method to transfer low-level filters from pretrained GANs and proposes AdaFM for domain adaptation with limited data, enhancing generation quality.

Findings

Transferred filters improve generation in limited data scenarios

AdaFM effectively adapts pretrained models to new domains

Experiments show significant quality improvements

Abstract

Recent work has shown generative adversarial networks (GANs) can generate highly realistic images, that are often indistinguishable (by humans) from real images. Most images so generated are not contained in the training dataset, suggesting potential for augmenting training sets with GAN-generated data. While this scenario is of particular relevance when there are limited data available, there is still the issue of training the GAN itself based on that limited data. To facilitate this, we leverage existing GAN models pretrained on large-scale datasets (like ImageNet) to introduce additional knowledge (which may not exist within the limited data), following the concept of transfer learning. Demonstrated by natural-image generation, we reveal that low-level filters (those close to observations) of both the generator and discriminator of pretrained GANs can be transferred to facilitate generation in a perceptually-distinct target domain with limited training data. To further adapt the transferred filters to the target domain, we propose adaptive filter modulation (AdaFM). An extensive set of experiments is presented to demonstrate the effectiveness of the proposed techniques on generation with limited data.