Sep 11 – 16, 2022
University of Warsaw
Europe/Warsaw timezone

3D scanning of celtic coins and Deep Learning to identify monetary dies

Sep 13, 2022, 3:40 PM
20m
Auditorium Maximum - Hall B

Auditorium Maximum - Hall B

Speakers

François Goulette (Mines Paris, PSL University) Jean-Emmanuel Deschaud (Mines Paris, PSL University) Katherine Gruel (Aoroc _CNRS- ENS-PSL) Sofiane Horache (Mines Paris, PSL University) Thierry Lejars (ENS Ulm, PSL University)

Description

To cluster thousands of coins, automatic methods are necessary. Public datasets for coin die clustering evaluation are too rare, despite their importance for the development of new methods using Artificial Intelligence. Therefore, with our dataset of 2070 3D scans of coins, we create two benchmarks, one for point cloud registration, essential for coin die recognition, and a benchmark of coin die clustering. We show how we automatically built the dataset and perform a preliminary evaluation. The code of the baseline and the dataset are publicly available. These results have been obtained by new developments using 3D imaging and the registration of 3D images by deep learning methods. We will present the die recognition algorithm with a visualisation tool with the possibility of excluding links between two coins that are rejected, and obtain an automatic redistribution of sets of coins from the same die.

Primary author

Jean-Emmanuel Deschaud (Mines Paris, PSL University)

Co-authors

François Goulette (Mines Paris, PSL University) Katherine Gruel (Aoroc _CNRS- ENS-PSL) Sofiane Horache (Mines Paris, PSL University) Thierry Lejars (ENS Ulm, PSL University)

Presentation materials

There are no materials yet.