MSc Dissertation, Abertay University

Improving Procedural Terrain Generation using Machine Learning

Scroll

Procedural Terrain Generation with ML

Image

Image

Image

Image

Image

Image

Project Description

A machine learning(ML) agent is trained to improve the noise parameters used for terrain. The concept of using an ML agent to improve PCG showcased here can be used to complement the work of designers or to set a baseline that the designers can then modify.

Shader Graph is used to create height based, slope based terrain and also physical water waves.

Project Date July, 2020 to Aug 2020 (2 months)
Snow
Forest