Analysis of PDE-based binarization model for degraded document images

TL;DR

This paper evaluates a PDE-based binarization model for degraded document images, demonstrating its effectiveness in handling bleed-through, faded text, and stains, thereby improving document image processing.

Contribution

It introduces a PDE-based binarization approach utilizing edge and binary source terms tailored for degraded documents, showing improved performance over existing methods.

Findings

Effective in binarizing images with bleed-through and faded text

Less effective for images with stains

Demonstrates potential for improved document image processing

Abstract

This report presents the results of a PDE-based binarization model for degraded document images. The model utilizes an edge and binary source term in its formulation. Results indicate effectiveness for document images with bleed-through and faded text and stains to a lesser extent.