CVMay 2, 2019

SinGAN: Learning a Generative Model from a Single Natural Image

arXiv:1905.01164v2938 citations
Originality Incremental advance
AI Analysis

This enables image manipulation tasks from a single image, but it is incremental as it builds on GANs for single-image generation.

The paper tackles the problem of learning a generative model from a single natural image, resulting in SinGAN, which generates high-quality, diverse samples that maintain the visual content of the original image, with user studies showing confusion with real images.

We introduce SinGAN, an unconditional generative model that can be learned from a single natural image. Our model is trained to capture the internal distribution of patches within the image, and is then able to generate high quality, diverse samples that carry the same visual content as the image. SinGAN contains a pyramid of fully convolutional GANs, each responsible for learning the patch distribution at a different scale of the image. This allows generating new samples of arbitrary size and aspect ratio, that have significant variability, yet maintain both the global structure and the fine textures of the training image. In contrast to previous single image GAN schemes, our approach is not limited to texture images, and is not conditional (i.e. it generates samples from noise). User studies confirm that the generated samples are commonly confused to be real images. We illustrate the utility of SinGAN in a wide range of image manipulation tasks.

Code Implementations47 repos

Data from Papers with Code (CC-BY-SA-4.0)

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes