Tuesday, September 24, 1:30 PM – 4:30 PM
Led by Peter Leonard and Damon Crocket
Pre-trained neural nets are widely available and offer new opportunities for cultural heritage image classification. Libraries and archives, often in the context of digital humanities, are making use of existing data sets of digitized images to propose new modes of accessing and understanding large photograph collections. This seminar will cover the basics of neural nets as well as offering a case study for neural net classification from the Meserve-Kunhardt Collection of civil war era photographs jointly held by Yale’s Beinecke Rare Book & Manuscript Library and the Yale University Art Gallery. The seminar will provoke discussion on how these emerging techniques can be made accessible and refined to address research questions for curators, art historians, and conservators.