ShabbyPages: A Reproducible Document Denoising and Binarization Dataset

TL;DR

ShabbyPages is a new large-scale dataset for training and benchmarking document denoising and binarization models, featuring synthetically-noised images that simulate real-world document degradation.

Contribution

We introduce ShabbyPages, a comprehensive dataset with over 6,000 images, enabling improved training and evaluation of document denoising and binarization algorithms.

Findings

Convolutional denoisers trained on ShabbyPages effectively remove real noise.

ShabbyPages provides a challenging benchmark for future research.

Baseline models demonstrate high perceptual fidelity in noise removal.

Abstract

Document denoising and binarization are fundamental problems in the document processing space, but current datasets are often too small and lack sufficient complexity to effectively train and benchmark modern data-driven machine learning models. To fill this gap, we introduce ShabbyPages, a new document image dataset designed for training and benchmarking document denoisers and binarizers. ShabbyPages contains over 6,000 clean "born digital" images with synthetically-noised counterparts ("shabby pages") that were augmented using the Augraphy document augmentation tool to appear as if they have been printed and faxed, photocopied, or otherwise altered through physical processes. In this paper, we discuss the creation process of ShabbyPages and demonstrate the utility of ShabbyPages by training convolutional denoisers which remove real noise features with a high degree of human-perceptible fidelity, establishing baseline performance for a new ShabbyPages benchmark.