Nuke CopyCat nominated for VES Award

Developed to meet the needs of compositors and supervisors in a climate of increasing shot volume and complexity, Nuke’s CopyCat is a machine learning toolset. The technology is designed to equip artists with the tools they need to train their own machine learning model, with their own data, to create effects specific to the shots they’re working on.

The nomination highlighted the use of CopyCat on a large scale in the production of Dune: Part Two, where training data from Dune: Part One was used, to recreate the Fremen’s distinctive blue eyes. This submission was made possible thanks to the support of the team at Warner Bros. Pictures, plus Production VFX Supervisor, Paul Lambert and the compositing team at Wylie Co.

Showcasing the true potential of machine learning in the VFX industry, CopyCat enabled the Dune: Part Two team to reuse the manual work from the first film to train their custom model and re-color the eyes in a fraction of the time on the second film.

The nominated team from Foundry comprises:

• Ben Kent, Director of Research
• Guillaume Gales, Principal Research Engineer
• Mairead Grogan, Senior Research Engineer
• Johanna Barbier, Senior Research Engineer

The Emerging Technology Award was won by the Neural Performance Toolset on Here, while the other nominees included the Artist-driven Machine Learning Character on Furiosa: A Mad Max Saga, the Real-Time Interactive Filmmaking, From Stage To Post on Mufasa: The Lion King, and the Phase Synced Flash-Gun System on The Penguin.

Find out more about Nuke’s CopyCat machine learning toolset.