FUNSD: A Dataset for Form Understanding in Noisy Scanned Documents
This dataset addresses the problem of extracting structured information from noisy scanned forms for researchers in document analysis, though it is incremental as it builds on existing form understanding tasks.
The authors introduced FUNSD, a dataset of 199 annotated noisy scanned forms to tackle form understanding, providing baselines and evaluation metrics for tasks like text detection and entity linking.
We present a new dataset for form understanding in noisy scanned documents (FUNSD) that aims at extracting and structuring the textual content of forms. The dataset comprises 199 real, fully annotated, scanned forms. The documents are noisy and vary widely in appearance, making form understanding (FoUn) a challenging task. The proposed dataset can be used for various tasks, including text detection, optical character recognition, spatial layout analysis, and entity labeling/linking. To the best of our knowledge, this is the first publicly available dataset with comprehensive annotations to address FoUn task. We also present a set of baselines and introduce metrics to evaluate performance on the FUNSD dataset, which can be downloaded at https://guillaumejaume.github.io/FUNSD/.