Sketch-a-Map (SAM): Creative Route Art Generation

Abstract

The creation of navigable, geographic art and its interactions is a novel, ideal environment for a computationally creative system. In this paper, we present a system that utilizes a canvas of city blocks to draw paths that represent the mood of the user, or if none is provided, it picks inspiration for itself. By being provided with user mood input, the system then identifies images that represent the provided emotion and, on its own volition, plots a path for the user to traverse. This kind of art is often classified as Strava Art which has been a hobby within online groups as early as 2015. The presented system utilizes techniques for translation and navigation of geodetic data provided by OpenStreetMap, processing images in the knowledge base for geographic display, and error checking through the use of Hausdorff calculations. Through a publicly administered survey, it was demonstrated that the system is effective at generating artifacts that are representative of the selected image. Source Code: https://github.com/epidermus/SAM