PT Unknown
AU V. Poulain d'Andecy
   Emmanuel Hartmann
   Marçal Rusiñol
TI Field Extraction by hybrid incremental and a-priori structural templates
BT 13th IAPR International Workshop on Document Analysis Systems
PY 2018
BP 251
EP 256
DI 10.1109/DAS.2018.29
DE Layout Analysis; information extraction; incremental learning
AB In this paper, we present an incremental framework for extracting information fields from administrative documents. First, we demonstrate some limits of the existing state-of-the-art methods such as the delay of the system efficiency. This is a concern in industrial context when we have only few samples of each document class. Based on this analysis, we propose a hybrid system combining incremental learning by means of itf-df statistics and a-priori genericmodels. We report in the experimental section our results obtained with a dataset of real invoices.
ER