.png){kind=icon}
.png){kind=icon}
Return
EPIMETHEE - A Workflow from OCR to Spatial Mapping
.png){kind=icon}
PRE
AI
DOI：10.1007/978-3-032-04624-6_1.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
We present the elaboration of EPIMETHEE, a text-processing pipeline that goes from Optical Character Recognition (OCR) to Named Entity Recognition (NER) and the cartographic representation of places mentioned in ancient literary texts. We will present the difficulties encountered when using off-the-shelf tools for the NER and Map stages in noisy data and the methods used to overcome them. One involves grouping different versions of NEs, for example Besancon, Besangon or besanqon, using a clustering algorithm. We present the assessment for several clustering algorithms. The analysis of spatial NEs with EPIMETHEE lead researchers to better understand, represent and deepen the stakes of a novel by observing the diegetic landscapes proposed by literary authors.
Keywords:
Named entity recognition
Optical character recognition
Noisy Documents
Robustness
Journal
D
IF:
0
Papers:
31
Citations:
0