Reference-Conditioned Super-Resolution by Neural Texture Transfer

TL;DR

This paper introduces a reference-conditioned super-resolution method that transfers high-resolution textures from reference images to enhance low-resolution images, demonstrating superior results over existing techniques.

Contribution

It proposes a neural texture transfer approach for super-resolution that does not require content similarity between reference and target images, along with a new benchmark dataset.

Findings

Outperforms state-of-the-art super-resolution methods

Effectively transfers high-resolution textures from references

Creates a new dataset for reference-based super-resolution

Abstract

With the recent advancement in deep learning, we have witnessed a great progress in single image super-resolution. However, due to the significant information loss of the image downscaling process, it has become extremely challenging to further advance the state-of-the-art, especially for large upscaling factors. This paper explores a new research direction in super resolution, called reference-conditioned super-resolution, in which a reference image containing desired high-resolution texture details is provided besides the low-resolution image. We focus on transferring the high-resolution texture from reference images to the super-resolution process without the constraint of content similarity between reference and target images, which is a key difference from previous example-based methods. Inspired by recent work on image stylization, we address the problem via neural texture transfer. We design an end-to-end trainable deep model which generates detail enriched results by adaptively fusing the content from the low-resolution image with the texture patterns from the reference image. We create a benchmark dataset for the general research of reference-based super-resolution, which contains reference images paired with low-resolution inputs with varying degrees of similarity. Both objective and subjective evaluations demonstrate the great potential of using reference images as well as the superiority of our results over other state-of-the-art methods.