On Leveraging Pretrained GANs for Generation with Limited Data
This addresses data scarcity in generative modeling for domains like natural images, but it is incremental as it builds on existing transfer learning concepts.
The paper tackles the problem of training GANs with limited data by leveraging pretrained GANs and introducing adaptive filter modulation (AdaFM) to transfer low-level filters to perceptually-distinct domains, achieving effective generation as demonstrated in experiments.
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.