Please go to the following link for the IASA - FIAT/IFTA Conference:

Back To Schedule
Tuesday, October 27 • 09:00 - 09:30
Studying representation on British television using computer vision and deep learning

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

There is a growing evidence base around diversity and equal opportunities in television. In the UK, the broadcasting regulator Ofcom has published the annual Diversity in Broadcasting report since 2017, requiring the main five UK broadcasters to provide data on legally protected characteristics such as gender, racial group, and disability of workers, both on and off screen.

Nesta (an innovation foundation) and Learning on Screen (who operate BoB, an on demand TV and radio service for education) have partnered to investigate how to apply novel methods to archived broadcast content for more efficient diversity monitoring.

Using a sample of broadcast content from free-to-air channels, we apply computer vision and machine learning methods (including face detection and facial attribute classification algorithms using deep learning) to analyse the characters who are shown on screen and how they are represented on British television. We discuss the findings and challenges of using these novel research methods, including those posed by copyright law. Getting permission from rights holders to use large amounts of copyright protected works for AI purposes is often impractical, if not impossible. In this context, can users of audiovisual archives rely on copyright exceptions? How can archivists assist and advise users who wish to use protected materials under exceptions? This parallel session presentation will explore these challenging questions as well as other ethical, socio-legal and technical implications.

avatar for Judith Opoku-Boateng

Judith Opoku-Boateng

Archivist, Institute of African Studies, University of Ghana

avatar for Raphael Leung

Raphael Leung

Data scientist, Nesta
Raphael Leung is a Principal Data Scientist at Nesta, where he uses novel data and data science to generate useful evidence in the form of peer-reviewed research outputs for the creative industries.Recently, using web-scraping and natural language processing, he analysed text on arts... Read More →
avatar for Bartolomeo Meletti

Bartolomeo Meletti

Education and Research Executive, Learning on Screen, UK
Bartolomeo Meletti is the Education and Research Executive of Learning on Screen, a membership organisation specialised in the use of moving image and sound in education. He also works as Creative Director for CREATe, the UK Copyright and Creative Economy Centre at the University... Read More →

Tuesday October 27, 2020 09:00 - 09:30 GMT
d. Session 3