Saved in:
Bibliographic Details
Main Authors: Villanova-Aparisi, David, Tarride, Solène, Martínez-Hinarejos, Carlos-D., Romero, Verónica, Kermorvant, Christopher, Pastor-Gadea, Moisés
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2404.18664
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914774469574656
author Villanova-Aparisi, David
Tarride, Solène
Martínez-Hinarejos, Carlos-D.
Romero, Verónica
Kermorvant, Christopher
Pastor-Gadea, Moisés
author_facet Villanova-Aparisi, David
Tarride, Solène
Martínez-Hinarejos, Carlos-D.
Romero, Verónica
Kermorvant, Christopher
Pastor-Gadea, Moisés
contents Information Extraction processes in handwritten documents tend to rely on obtaining an automatic transcription and performing Named Entity Recognition (NER) over such transcription. For this reason, in publicly available datasets, the performance of the systems is usually evaluated with metrics particular to each dataset. Moreover, most of the metrics employed are sensitive to reading order errors. Therefore, they do not reflect the expected final application of the system and introduce biases in more complex documents. In this paper, we propose and publicly release a set of reading order independent metrics tailored to Information Extraction evaluation in handwritten documents. In our experimentation, we perform an in-depth analysis of the behavior of the metrics to recommend what we consider to be the minimal set of metrics to evaluate a task correctly.
format Preprint
id arxiv_https___arxiv_org_abs_2404_18664
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Reading Order Independent Metrics for Information Extraction in Handwritten Documents
Villanova-Aparisi, David
Tarride, Solène
Martínez-Hinarejos, Carlos-D.
Romero, Verónica
Kermorvant, Christopher
Pastor-Gadea, Moisés
Computer Vision and Pattern Recognition
Information Extraction processes in handwritten documents tend to rely on obtaining an automatic transcription and performing Named Entity Recognition (NER) over such transcription. For this reason, in publicly available datasets, the performance of the systems is usually evaluated with metrics particular to each dataset. Moreover, most of the metrics employed are sensitive to reading order errors. Therefore, they do not reflect the expected final application of the system and introduce biases in more complex documents. In this paper, we propose and publicly release a set of reading order independent metrics tailored to Information Extraction evaluation in handwritten documents. In our experimentation, we perform an in-depth analysis of the behavior of the metrics to recommend what we consider to be the minimal set of metrics to evaluate a task correctly.
title Reading Order Independent Metrics for Information Extraction in Handwritten Documents
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2404.18664