Efficient Cost-and-Quality Controllable Arbitrary-scale Super-resolution with Fourier Constraints

TL;DR

This paper introduces a novel super-resolution method that predicts multiple Fourier components simultaneously, enhancing both the quality and efficiency of arbitrary-scale super-resolution with controllable cost and quality.

Contribution

It proposes a joint prediction approach for Fourier components in super-resolution, overcoming the limitations of previous independent prediction methods.

Findings

Improved super-resolution quality and efficiency.

Effective control over cost and quality in arbitrary-scale super-resolution.

Demonstrated superiority over existing methods.

Abstract

Cost-and-Quality (CQ) controllability in arbitrary-scale super-resolution is crucial. Existing methods predict Fourier components one by one using a recurrent neural network. However, this approach leads to performance degradation and inefficiency due to independent prediction. This paper proposes predicting multiple components jointly to improve both quality and efficiency.