Skip to main content
School of Electronic Engineering and Computer Science

Breadcrumb

  1. School of Electronic Engineering and Computer Science
  2. News & Events
  3. Events
  4. Items

Game AI Seminar - Gabriel Bosque: Cinematic animation in RPG

When: Wednesday, November 30, 2022, 1:00 PM - 2:00 PM
Where: Online and in-person at Empire House

Speaker: Gabriel Bosque

Cinematics are fundamental to any AAA RPG game. They move the story forward and give players truly cinematic moments. These sequences can range from simple conversations to large action set pieces. Cinematic animation deals with producing high quality and human-like movement for those sequences. A widely used method to get these animations is motion capture, where an actor is recorded moving in a space while wearing a suit and a set of markers, that are later translated into character animation. Part of the challenge in motion capture comes from the inherent shortcomings of the process, that produces errors such as jitter and swapped markers. Machine Learning can help with cleaning this data, and thus improving the efficiency of the entire pipeline.
On the other hand, facial animation is absolutely key for an immersive player experience during cinematics. There are different approaches to produce realistic facial animation, such as using facial emotion recognition on the actors when the motion capture is recorded. How to use such a system and make it available for cinematic artists to compose scenes can be challenging. Gabriel will cover how his team approached that issue at Larian.

Zoom joining link

https://qmul-ac-uk.zoom.us/j/83945925780 

Speaker biography

Gabriel Bosque is the Machine Learning Lead at Larian Studios, a AAA developer specialised in narrative heavy RPGs. Gabriel got his PhD at the Automatics and Industrial Informatics Institute at the Polytechnic University of Valencia. After that he has been working in tech start-ups like Deliveroo, Cabify or Lingokids as an Machine Learning Engineer and ML Lead. He recently joined Larian to create its new Machine Learning team, based in Guildford, UK.

Back to top