Affiliation: Associate Professor. Department of Computer Science. Grinnell College. IA, USA.
Place: Large Lecture Room
Recognizing the place names within textual labels on historical maps is complicated by many factors, such as curvilinear baselines and dense overlap with other textual or graphical elements. However, maps’ intrinsic alignment with known geography and inter-label typographic style consistencies provide strong cues for resolving uncertainty and reducing recognition errors. I will present a unified probabilistic model to leverage the mutual information between text labels and styles and their geographical locations and categories. I show how interleaving automated map georeferencing with text recognition reduces word recognition error by 36% and incorporating category-style links reduces toponym matching error by 32%.
Jerod Weinman is Associate Professor of Computer Science at Grinnell College (Iowa, USA), and currently visiting the CVC on sabbatical. He holds a Bachelor of Science in Computer Science and Mathematics from Rose-Hulman Institute of Technology and a Master of Science and PhD from the University of Massachusetts Amherst. He is a senior member of both the ACM and IEEE, as well as a member of ACM’s SIGCSE and SIGCAS.