Skip to Main content Skip to Navigation
Master Thesis

Humanités numériques : un cas pratique avec l'identification de hiéroglyphes

Abstract : As part of the VEgA project, we are interested in the use of image recognition methods using convolutional neural networks in the context of digital humanities. We show in particular how classification can be performed on hieroglyphs. We present our classification model and the graphical user interface and we analyse our experimental evaluation. Although we only have few training examples (around 44 hieroglyphs per class), our model achieves more than 90% accuracy on the validation set.
Complete list of metadatas

Cited literature [17 references]  Display  Hide  Download

https://dumas.ccsd.cnrs.fr/dumas-02097737
Contributor : Ufr6 Um3 <>
Submitted on : Friday, April 12, 2019 - 11:18:04 AM
Last modification on : Wednesday, October 7, 2020 - 10:04:31 AM

File

Madeleine_2018_UFR6.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : dumas-02097737, version 1

Collections

Citation

Sarah Madeleine, Sylvain Maillot, Sara Thomas. Humanités numériques : un cas pratique avec l'identification de hiéroglyphes. Archéologie et Préhistoire. 2018. ⟨dumas-02097737⟩

Share

Metrics

Record views

45

Files downloads

36