Enhancing visual depictions on digital screens through cutting-edge computer vision techniques
The UK screen industry is undergoing a significant transformation, with a new scoping study exploring the economic consequences and potential market failures of overseas mergers and acquisitions. This research, coupled with a report detailing the migrant and skills needs of creative businesses in the UK, paints a picture of an industry in flux.
Amidst these changes, the importance of diversity and representation within the screen industry has never been more crucial. To address this issue, a groundbreaking project led by Nesta and Learning on Screen is using computer vision to systematically measure on-screen diversity.
The project, inspired by research from Creative PEC, aims to provide a new method for quantifying representation. By leveraging advanced deep learning models and image analysis techniques, computer vision can identify and classify faces, demographics, poses, and actions in film and television content.
This approach allows for the quantification of diverse groups on-screen, as well as a richer analysis of representation beyond just presence. For instance, computer vision can determine who speaks, acts, or appears prominently, providing a more nuanced understanding of character representation.
Through aggregating such data, systems can compute diversity indices, reflecting how inclusive casting and representation are across productions. Moreover, platforms like Varif.ai propose empowering users to define and measure diversity in image sets and iteratively improve it based on computer vision feedback loops.
However, the use of computer vision for this purpose raises multiple ethical concerns. Privacy and consent, bias and fairness, interpretation and impact, and transparency and accountability are all critical factors that must be addressed to ensure responsible application.
Despite some progress, the UK screen industry continues to grapple with persistent diversity issues. The project team, including Raphael Leung, Bartolomeo Meletti, Dr Cath Sleeman, Gabriel A. Hernández, and Gil Toffell, is hopeful that computer vision can provide a valuable tool in addressing these challenges.
The work, which builds upon a presentation given at the International Federation of Television Archives (FIAT/IFTA) Conference 2020 and a short paper by Raphael for the IJCAI 2021 AI for Social Good workshop, is set to explore whether computer vision can provide a new, effective method for measuring on-screen representation.
In a global creative economy worth over half a trillion dollars, with a 150% increase since 2000, the UK screen industry stands to benefit significantly from addressing diversity issues. Post-Brexit, accessing foreign talent in the creative industries has become more complex, making it all the more important to ensure a diverse and inclusive industry at home.
- The new scoping study in the UK screen industry is examining the economic impacts of overseas mergers and acquisitions, as part of a broader industry transformation.
- A project led by Nesta and Learning on Screen is using computer vision to systematically measure on-screen diversity in the industry, addressing the persistent issues of diversity and representation.
- Advanced deep learning models and image analysis techniques are being employed to identify and classify demographics, poses, and actions in film and television content, providing a more nuanced understanding of character representation.
- Aggregated data from this approach allows for the computation of diversity indices, reflecting the inclusivity of casting and representation across productions.
- Platforms like Varif.ai enable users to define and measure diversity in image sets, with the goal of iteratively improving representation using computer vision feedback loops.
- However, the ethical concerns of privacy, consent, bias, fairness, interpretation, impact, transparency, and accountability must be carefully considered to ensure a responsible application of computer vision.
- In a creative economy worth over half a trillion dollars, with post-Brexit access to foreign talent becoming more complex, addressing diversity issues in the UK screen industry can provide significant benefits, fostering a diverse and inclusive domestic talent pool.
