System Description of CITlab's Recognition & Retrieval Engine for ICDAR2017 Competition on Information Extraction in Historical Handwritten Records

TL;DR

This paper describes a neural network-based system for extracting and retrieving specific information from historical handwritten records, achieving high accuracy and outperforming baselines in the ICDAR2017 competition.

Contribution

The paper introduces a recognition and retrieval system that combines neural networks with regular expressions for extracting data from handwritten records, advancing methods in historical document analysis.

Findings

High accuracy in inferring person names and data

Outperforms baseline methods in the competition

Effective use of neural networks and regular expressions

Abstract

We present a recognition and retrieval system for the ICDAR2017 Competition on Information Extraction in Historical Handwritten Records which successfully infers person names and other data from marriage records. The system extracts information from the line images with a high accuracy and outperforms the baseline. The optical model is based on Neural Networks. To infer the desired information, regular expressions are used to describe the set of feasible words sequences.