Cartographic Recognition: Using Machine Learning Methods to Understand Maps
Recent developments in machine learning have made it possible for computers to mimic or outperform humans in various recognitive applications. We explore if these methods can be used to recognize maps, an area we broadly refer to as cartographic recognition. Specifically, we develop machine learning methods to identify if an image is a map or not, then recognize the geographic region on the map image and the projection used in the map. Besides the identification tasks, we also test the limit of the methods to recognize the information where the maps are changed in some aspects including shape and position.
Zoom link: https://osu.zoom.us/j/95483016042