This was the final course project for an Intro to Digital and Computational Studies course at Bowdoin College in Fall 2017.
Our class worked in collaboration with the Bowdoin College Museum of Art to facilitate the digitization of the Molinari medal collection, one of the largest collections in North America. Cesare Molinari's personal collection of 1500+ medals and plaquettes dating from 1438 onward were donated to Bowdoin College in the 1960s and a catalogue detailing 357 of these was published online in 2015. The Museum of Art aims to fully digitize the 805 medals which have been accessioned, and to create a searchable database with images of the medals. Our class hand-tagged a selection of these medals and used TensorFlow to train a machine learning model to tag photos of the medals. This will be indispensable for allowing users to search for medals by category and obtain accurate results.
The primary code and instructions to run the model were provided by our professor, though we were asked to edit parts of it to ensure that we understood it.
Bowdoin College Museum of Art Molinari Medals Collection Catalogue: https://archive.org/details/medalsplaquettes00bowd