Generating Stylised Game Assets using Generative Adversarial Networks
The aim of this project is to investigate how Generative Adversarial Networks (GANs) can allow the creation of detailed, beautiful backgrounds to increase a game’s narrative ambition by providing larger, more extensive environments through which stories can unfold.
Unlike a linear medium like the cartoon, an adventure game invites the player to investigate and search for clues in the intricate environments. As these stories become increasingly ambitious, so does the requirement for increasingly visually appealing as well as detailed backgrounds to support and enhance them. Audiences are increasingly choosing to purchase the most visually stunning and narratively rich adventure games.
To date, the backgrounds for adventure games have required the creation of huge volumes of art assets in a very labour-intensive process. The ambition of these stories and adventures are heavily limited by the production cost of 2D/3D art assets. As games move from 2D to full 3D, the issue with retaining the hand-drawn visual look is exacerbated.
This summer school project will build on the work of DC Labs researchers who worked with Revolution Software last year to accelerate the generation of layouts for adventure games.
The summer school student would extend this work in the following ways:
- Run the existing GAN code on a number of artist style/target images to produce generated textures in the style of the artist.
- Perform a parameter tuning/sweep investigation using the existing code to assess the impact of each parameter on the resulting stylised designs and overall performance of the GAN.
We are seeking a candidate for one role, with the following spread of skills:
- Python programming (Essential)
- Knowledge of GANs/Deep Learning (Essential)
- Interest in digital art (Essential)
How to Apply
For more details on the summer school application process (including eligibility and funding) please go here: https://digitalcreativity.ac.uk/news/dc-labs-summer-school-2021
For questions about the programme, logistics etc, please contact Ella Eyre, DC Labs Administrator, on firstname.lastname@example.org