A Reinforcement Learning Approach to Determine Horizontal Spaces in Typefaces

Abstract

Abstract Typeface spacing is a hard problem. It takes countless hours of manual labour to achieve an aesthetically pleasing font, one frequently encounters in digital media. The amount of space between two letters (inter-letter space) significantly contributes to the aesthetically pleasing nature and readability of the typeface. Although inter-letter spacing defines the texture and feel of a typeface and when done accurately yields aesthetically balanced, and an appealing typeface. Setting spacing in a typeface is a tedious and time-consuming task. Hence this research presents an exploratory study investigating potential of reinforcement learning models to fully automate the typeface spacing process. The proposed reinforcement learning model, first of it’s kind was able to achieve good accuracies even with a simple reward function. Some of the visual differences were subtle. Thus, we conclude that reinforcement learning models can indeed be used to model the typeface spacing problem and as one of the first attempts to apply reinforcement learning models in this particular problem domain, this study lays the foundation to future research and studies.