tile2tile: Learning Game Filters for Platformer Style Transfer

TL;DR

tile2tile introduces a novel method for style transfer in tile-based platformer games by translating level sketches into game-specific tiles, enabling cross-game level style transfer using learned filters.

Contribution

The paper presents a new approach using Markov random fields and autoencoders to perform style transfer between different platformer game levels.

Findings

Successful style transfer demonstrated between multiple classic games

Models effectively translate level sketches into game-specific tiles

Enables cross-game level style adaptation

Abstract

We present tile2tile, an approach for style transfer between levels of tile-based platformer games. Our method involves training models that translate levels from a lower-resolution sketch representation based on tile affordances to the original tile representation for a given game. This enables these models, which we refer to as filters, to translate level sketches into the style of a specific game. Moreover, by converting a level of one game into sketch form and then translating the resulting sketch into the tiles of another game, we obtain a method of style transfer between two games. We use Markov random fields and autoencoders for learning the game filters and apply them to demonstrate style transfer between levels of Super Mario Bros, Kid Icarus, Mega Man and Metroid.