Agenda
The Scalable Game Asset-Assisted Production Pipeline of 'Honor of Kings'
Driven by the demand for immersive and visually stunning experiences, developers face the challenge of creating detailed assets efficiently and cost-effectively across multiple platforms. Traditional asset creation methods are time-consuming and expensive, making them inadequate for this evolving landscape.
This presentation explores how AI techniques, such as differentiable rendering and Artificial Intelligence Generated Content (AIGC), enhance asset production quality and efficiency. Specifically, within our game Honor Of Kings and other titles, we utilize Differential Shading Language (Slang) for material fitting, optimizing textures for low-end devices, and incorporating tools for artists to swiftly iterate designs. Innovations in geometry processing, including Software Occlusion Culling and Level of Detail (LOD) simplification, leverage rendering loss and visual importance to maintain high-quality visual experiences while reducing resource consumption.
By adopting these cutting-edge AI technologies, we significantly improve the creation and deployment of game assets, addressing the diverse challenges of modern game development.
Takeaway
Attendees will gain insight into how AI technology can be employed to power game asset production for both improving the quality and reducing time consumption to generate high-detail or low-end devices runnable game assets, in several specific domains like material making, geometry production, etc.
Intended Audience
This presentation is intended for developers who want to employ AI technology to power their game asset production pipelines, as well as engineers who are interested in implementing a differentiable rendering pipeline in game engine. Cursory knowledge of rendering and artificial intelligence is recommended.