Progressive Image Super-Resolution via Neural Differential Equation

TL;DR

This paper introduces a novel neural differential equation-based approach for progressive image super-resolution, enabling arbitrary scale factor super-resolution with improved performance and clearer modeling compared to traditional methods.

Contribution

The paper formulates the super-resolution task as an initial value problem using neural ordinary differential equations, providing a more explicit and flexible framework for progressive image restoration.

Findings

Achieves superior super-resolution performance over state-of-the-art methods.

Supports arbitrary scale factors in continuous domain.

Provides a clearer, more explicit modeling approach for progressive SR.

Abstract

We propose a new approach for the image super-resolution (SR) task that progressively restores a high-resolution (HR) image from an input low-resolution (LR) image on the basis of a neural ordinary differential equation. In particular, we newly formulate the SR problem as an initial value problem, where the initial value is the input LR image. Unlike conventional progressive SR methods that perform gradual updates using straightforward iterative mechanisms, our SR process is formulated in a concrete manner based on explicit modeling with a much clearer understanding. Our method can be easily implemented using conventional neural networks for image restoration. Moreover, the proposed method can super-resolve an image with arbitrary scale factors on continuous domain, and achieves superior SR performance over state-of-the-art SR methods.