Handwritten Text Recognition from Crowdsourced Annotations
This addresses the challenge of handling imperfect annotations in historical document digitization, but it is incremental as it builds on existing methods for noisy data.
The paper tackles the problem of training handwritten text recognition models using multiple noisy transcriptions, finding that computing a consensus or using all transcriptions works well, but quality-based data selection introduces bias and does not improve results, with experiments conducted on historical municipal registers from 1790-1946.
In this paper, we explore different ways of training a model for handwritten text recognition when multiple imperfect or noisy transcriptions are available. We consider various training configurations, such as selecting a single transcription, retaining all transcriptions, or computing an aggregated transcription from all available annotations. In addition, we evaluate the impact of quality-based data selection, where samples with low agreement are removed from the training set. Our experiments are carried out on municipal registers of the city of Belfort (France) written between 1790 and 1946. % results The results show that computing a consensus transcription or training on multiple transcriptions are good alternatives. However, selecting training samples based on the degree of agreement between annotators introduces a bias in the training data and does not improve the results. Our dataset is publicly available on Zenodo: https://zenodo.org/record/8041668.