End-to-End Unsupervised Document Image Blind Denoising

TL;DR

This paper introduces a novel unsupervised deep learning model capable of removing multiple noise types from scanned document images, significantly enhancing image quality and OCR accuracy without requiring noisy-clean image pairs.

Contribution

It presents the first unified end-to-end unsupervised model for removing various noise types from document images, addressing a key limitation of prior supervised methods.

Findings

Improves scanned image quality across multiple noise types.

Enhances OCR accuracy on test datasets.

Operates effectively without paired training data.

Abstract

Removing noise from scanned pages is a vital step before their submission to the optical character recognition (OCR) system. Most available image denoising methods are supervised where the pairs of noisy/clean pages are required. However, this assumption is rarely met in real settings. Besides, there is no single model that can remove various noise types from documents. Here, we propose a unified end-to-end unsupervised deep learning model, for the first time, that can effectively remove multiple types of noise, including salt \& pepper noise, blurred and/or faded text, as well as watermarks from documents at various levels of intensity. We demonstrate that the proposed model significantly improves the quality of scanned images and the OCR of the pages on several test datasets.