Deblurring Photographs of Characters Using Deep Neural Networks

TL;DR

This paper introduces a deep learning-based approach to deblur images of characters without prior knowledge of the PSF, successfully reconstructing images in the Helsinki Deblur Challenge 2021.

Contribution

It presents a novel three-step method combining warping, PSF estimation, and neural network training for blind image deblurring.

Findings

Successfully reconstructed images from the first 10 stages of HDC 2021 data.

Demonstrated effectiveness of combined warping and PSF estimation in deblurring.

Code implementation available for reproducibility.

Abstract

In this paper, we present our approach for the Helsinki Deblur Challenge (HDC2021). The task of this challenge is to deblur images of characters without knowing the point spread function (PSF). The organizers provided a dataset of pairs of sharp and blurred images. Our method consists of three steps: First, we estimate a warping transformation of the images to align the sharp images with the blurred ones. Next, we estimate the PSF using a quasi-Newton method. The estimated PSF allows to generate additional pairs of sharp and blurred images. Finally, we train a deep convolutional neural network to reconstruct the sharp images from the blurred images. Our method is able to successfully reconstruct images from the first 10 stages of the HDC 2021 data. Our code is available at https://github.com/hhu-machine-learning/hdc2021-psfnn.