Computer vision for Digital Humanists: a critical and practical introduction

Machine vision systems can potentially help humanities researchers see historical and cultural image collections differently, and could provide tools to answer new research questions. This session provides an introductory overview of basic tasks in machine vision, such as Image Matching, Comparison, Classification, Captioning, Object Detection and Facial Recognition within a critical framework highlighting the challenges of algorithmic bias and the limits of automation as a method for humanistic enquiry. Students will have the opportunity to explore some demos developed by the Visual Geometry Group at the University of Oxford and create a small image-based Machine Learning project using Google’s Teachable Machine platform which will explore the challenges of applying machine vision to historical document collections. No prior knowledge of computer vision or of coding is required to follow the session.


This training comprises of two live sessions with time for self-directed learning in between. Session 1 will take place from 2:00 - 3:00pm, and session 2 will run from 4:00pm - 4:30pm. The time between the sessions is a required part of the course where students will be working on a small project of their own which they will present and discuss in the final live session. 


Please register to attend via Inkpath here, by 18 May 2022.