Recovering Realistic Texture in Image Super-resolution by Deep Spatial Feature Transform

TL;DR

This paper introduces a novel Spatial Feature Transform layer that modulates CNN features based on semantic segmentation maps, significantly improving the realism of textures in image super-resolution tasks.

Contribution

The paper proposes the SFT layer for spatial-wise feature modulation conditioned on semantic priors, enabling more realistic texture recovery in super-resolution.

Findings

SFT improves texture realism over state-of-the-art methods.

The method works with arbitrary input sizes in a single pass.

Enhanced visual quality in super-resolved images.

Abstract

Despite that convolutional neural networks (CNN) have recently demonstrated high-quality reconstruction for single-image super-resolution (SR), recovering natural and realistic texture remains a challenging problem. In this paper, we show that it is possible to recover textures faithful to semantic classes. In particular, we only need to modulate features of a few intermediate layers in a single network conditioned on semantic segmentation probability maps. This is made possible through a novel Spatial Feature Transform (SFT) layer that generates affine transformation parameters for spatial-wise feature modulation. SFT layers can be trained end-to-end together with the SR network using the same loss function. During testing, it accepts an input image of arbitrary size and generates a high-resolution image with just a single forward pass conditioned on the categorical priors. Our final results show that an SR network equipped with SFT can generate more realistic and visually pleasing textures in comparison to state-of-the-art SRGAN and EnhanceNet.