Record Counting in Historical Handwritten Documents with Convolutional Neural Networks

TL;DR

This paper demonstrates that convolutional neural networks trained solely on synthetic images can accurately count records in historical handwritten documents, achieving near-perfect results on a benchmark dataset.

Contribution

The study introduces a method using CNNs trained exclusively on synthetic data for record counting in historical documents, surpassing previous approaches.

Findings

Achieved near-perfect record counting accuracy on benchmark dataset.

Outperformed previous methods on the marriage records dataset.

Validated the effectiveness of synthetic training data for historical document analysis.

Abstract

In this paper, we investigate the use of Convolutional Neural Networks for counting the number of records in historical handwritten documents. With this work we demonstrate that training the networks only with synthetic images allows us to perform a near perfect evaluation of the number of records printed on historical documents. The experiments have been performed on a benchmark dataset composed by marriage records and outperform previous results on this dataset.