A Frequency Domain Neural Network for Fast Image Super-resolution

TL;DR

This paper introduces a frequency domain neural network for image super-resolution that significantly accelerates processing while maintaining comparable quality, leveraging the convolution theorem and Hartley transform for efficiency.

Contribution

The paper proposes a novel frequency domain neural network architecture that improves computational efficiency and training simplicity for image super-resolution.

Findings

Network is 10-100 times faster than existing methods.

Maintains comparable super-resolution quality with minimal performance loss.

Uses Hartley transform to avoid complex numbers in frequency domain.

Abstract

In this paper, we present a frequency domain neural network for image super-resolution. The network employs the convolution theorem so as to cast convolutions in the spatial domain as products in the frequency domain. Moreover, the non-linearity in deep nets, often achieved by a rectifier unit, is here cast as a convolution in the frequency domain. This not only yields a network which is very computationally efficient at testing but also one whose parameters can all be learnt accordingly. The network can be trained using back propagation and is devoid of complex numbers due to the use of the Hartley transform as an alternative to the Fourier transform. Moreover, the network is potentially applicable to other problems elsewhere in computer vision and image processing which are often cast in the frequency domain. We show results on super-resolution and compare against alternatives elsewhere in the literature. In our experiments, our network is one to two orders of magnitude faster than the alternatives with an imperceptible loss of performance.