Riedones3D: a celtic coin dataset

Riedones3D dataset


Clustering coins with respect to their die is an important component of numismatic research and crucial for understanding the economic history of tribes (especially when literary production does not exist, like in celtic culture).

Nevertheless, public datasets for coin die clustering are too rare, though they are very important for the development of new methods. 

Therefore, we propose a new dataset of 2207 3D scans of coins. With this dataset, we propose two benchmarks, one for pattern registration between point clouds of coins, essential for coin die recognition, and a benchmark of coin die clustering. 

The full Riedones3D dataset of 3D scans of coins can be downloaded here:    Link to download Riedones3D dataset

The same dataset can be downloaded entirely or by coins on NAKALA: http://s802828443.onlinehome.fr/NAKALA_API_2/frontend/index.php#

More details on the protocol to build the dataset and benchmarks are in the paper: https://arxiv.org/pdf/2109.15033.pdf



Riedones3D Registration Benchmark

Registration of the same die pattern between two coins


A part of the full dataset has been built to evaluate registration methods of patterns between scans of coins.

The Riedones3D_Registration_Benchmark is avalaible here: https://cloud.mines-paristech.fr/index.php/s/AVhOXp54IVGYLTM

Details are in the readme.md at the root of the .zip file.



Riedones3D Clustering Benchmark

Synthetic view of the clustering test set into dies


A part of the full dataset has been built to evaluate clustering methods to group coins into the same dies.

The Riedones3D_Clustering_Benchmark is avalaible here: https://cloud.mines-paristech.fr/index.php/s/Fhq8CJTrAOrMTAI

Details are in the readme.md at the root of the .zip file.

Github code to perform coin die recognition and die clustering like in our article: https://github.com/humanpose1/riedones3d


If you use this dataset, do not forget to cite our article:

booktitle = {Eurographics Workshop on Graphics and Cultural Heritage},
editor = {Hulusic, Vedad and Chalmers, Alan},
title = {Riedones3D: a Celtic Coin Dataset for Registration and Fine-grained Clustering},
author = {Horache, Sofiane and Deschaud, Jean-Emmanuel and Goulette, François and Gruel, Katherine and Lejars, Thierry and Masson, Olivier},
year = {2021},
publisher = {The Eurographics Association},
ISBN = {978-3-03868-141-0},
pages = {83-92},
DOI = {10.2312/gch.20211410}



1. Any use of one or more files contained in these archives must mention the associated data present in the metadata.xlsx file

2. Any use of one or more files must observe the terms of the Creative Commons license: CC-BY-NC-ND-3.0

For more information, see: https://creativecommons.org/licenses/?lang=fr-FR


Institutional partners: