Image-to-Level: Generation and Repair


  • Eugene Chen University of Alberta
  • Christoph Sydora University of Alberta
  • Brad Burega University of Alberta
  • Anmol Mahajan University of Alberta
  • Abdullah University of Alberta
  • Matthew Gallivan University of Alberta
  • Matthew Guzdial University of Alberta



Procedural content generation via machine learning (PCGML) has recently gained research attention due to its ability to generate new game content with minimal user input. However, thus far those without machine learning expertise have been largely unable to use PCGML to generate content to fit their needs. This paper proposes the use of images as the input for a PCGML process to generate game levels. Intuitively, a user can submit an image, with the system returning the closest valid Super Mario Bros. game level. Our results indicate that at least for domains like Super Mario Bros. we can recreate a target level with high fidelity.




How to Cite

Chen, E., Sydora, C., Burega, B., Mahajan, A., Abdullah, A., Gallivan, M., & Guzdial, M. (2020). Image-to-Level: Generation and Repair. Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, 16(1), 189-195.