Recognizing Handwriting Styles in a Historical Scanned Document Using Unsupervised Fuzzy Clustering

TL;DR

This paper presents an unsupervised fuzzy clustering approach combined with PCA to identify handwriting styles and hand shifts in historical scanned documents, advancing forensic writer attribution without labeled data.

Contribution

It introduces a novel unsupervised method using fuzzy soft clustering and PCA for handwriting style recognition in historical documents, bypassing the need for labeled training data.

Findings

Successfully detected hand shifts in historical manuscripts.

Demonstrated effectiveness of unsupervised clustering for writer attribution.

Enhanced forensic analysis capabilities for digitized historical documents.

Abstract

The forensic attribution of the handwriting in a digitized document to multiple scribes is a challenging problem of high dimensionality. Unique handwriting styles may be dissimilar in a blend of several factors including character size, stroke width, loops, ductus, slant angles, and cursive ligatures. Previous work on labeled data with Hidden Markov models, support vector machines, and semi-supervised recurrent neural networks have provided moderate to high success. In this study, we successfully detect hand shifts in a historical manuscript through fuzzy soft clustering in combination with linear principal component analysis. This advance demonstrates the successful deployment of unsupervised methods for writer attribution of historical documents and forensic document analysis.