%0 Conference Proceedings
%T DocILE Benchmark for Document Information Localization and Extraction
%A Stepan Simsa
%A Milan Sulc
%A Michal Uricar
%A Yash Patel
%A Ahmed Hamdi
%A Matej Kocian
%A Matyas Skalicky
%A Jiri Matas
%A Antoine Doucet
%A Mickael Coustaty
%A Dimosthenis Karatzas
%B 17th International Conference on Document Analysis and Recognition
%D 2023
%V 14188
%F Stepan Simsa2023
%O DAG
%O exported from refbase (http://refbase.cvc.uab.es/show.php?record=3903), last updated on Wed, 17 Jan 2024 13:23:59 +0100
%X This paper introduces the DocILE benchmark with the largest dataset of business documents for the tasks of Key Information Localization and Extraction and Line Item Recognition. It contains 6.7k annotated business documents, 100k synthetically generated documents, and nearly 1M unlabeled documents for unsupervised pre-training. The dataset has been built with knowledge of domain- and task-specific aspects, resulting in the following key features: (i) annotations in 55 classes, which surpasses the granularity of previously published key information extraction datasets by a large margin; (ii) Line Item Recognition represents a highly practical information extraction task, where key information has to be assigned to items in a table; (iii) documents come from numerous layouts and the test set includes zero- and few-shot cases as well as layouts commonly seen in the training set. The benchmark comes with several baselines, including RoBERTa, LayoutLMv3 and DETR-based Table Transformer; applied to both tasks of the DocILE benchmark, with results shared in this paper, offering a quick starting point for future work. The dataset, baselines and supplementary material are available at https://github.com/rossumai/docile.
%K Document AI
%K Information Extraction
%K Line Item Recognition
%K Business Documents
%K Intelligent Document Processing
%U https://link.springer.com/chapter/10.1007/978-3-031-41679-8_9
%U http://refbase.cvc.uab.es/files/SSU2023.pdf
%P 147–166