Abstract
This paper proposes a tool for 3D Procedural Terrain Generation in games. The main objective of this study is to assess the quality of the generated terrains and thus the experience of the user. The tool measures how traversable a terrain is, how long it takes to walk across and how much of it is accessible. The produced results can be used to further refine the generation method. The evaluation consists of multiple terrains being produced, based on Generative Adversarial Networks (GANs) trained on satellite images. A path finding algorithm is then used to evaluate the traversability of each terrain between two points on the map. Results show that the proposed method can generate quite diverse and interesting maps and that the proposed metric can be used for evaluation of playability.
Original language | English |
---|---|
Title of host publication | An Explorative Design Process for Game Map Generation Based on Satellite Images and Playability Factors |
Number of pages | 4 |
Publisher | Association for Computing Machinery |
Publication date | 15 Sept 2020 |
Pages | 1-4 |
Article number | 72 |
ISBN (Electronic) | 978-1-4503-8807-8 |
DOIs | |
Publication status | Published - 15 Sept 2020 |
Event | FDG '20: International Conference on the Foundations of Digital Games - Bugibba, Malta Duration: 15 Sept 2020 → 18 Sept 2020 http://fdg2020.org/ |
Conference
Conference | FDG '20: International Conference on the Foundations of Digital Games |
---|---|
Country/Territory | Malta |
City | Bugibba |
Period | 15/09/2020 → 18/09/2020 |
Internet address |
Keywords
- generative adversarial networks
- map generation
- pathfinding
- procedural generation