Old Stones, New Technologies: Computer Vision and Medieval Walls

Old Stones, New Technologies: Computer Vision and Medieval Walls
2012 to 2013
Undergraduate Research

This project is part of a multi-year research and teaching initiative that will result in independent research and senior distinction theses for undergraduates.  A team of Duke undergraduate students is working closely with Professor Caroline Bruzelius (Art, Art History & Visual Studies) and Professor Carlo Tomasi (Computer Science) to collect data, develop and eventually test the new analytic systems with the intention of creating a systematic protocol for the study of walls, carved surfaces (flat and curved) and masonry construction in historic buildings and sculpture.  

Professor Bruzelius led the student team on a research trip to Naples over Spring break 2013.  They used the opportunity to test a new data capture system for use with medieval masonry.  They worked primarily in the church of San Lorenzo, a Franciscan basilica in the heart of medieval Naples. 

The students are experimenting with an analytic system for the study of historic buildings through pattern recognition, data mining, and texture analysis.  Their research works with computational analytics to examine the shapes, textures, materials, sizes, and colors of the stones used in medieval structures in order to extract information on the technology of stonecutting, and possibly identify the work of individual masons (tool marks are like signatures), as well as potentially provide educated estimates on the size of the labor force. 

As it appears that work of this nature is not being done elsewhere, these Duke students may be in a position to provide an original contribution to scholarship of on-going utility that might have broader implications for fields of study in restoration of historic monuments as well as Ancient and Medieval Sculpture, Archaeology, and Architectural and Urban History.


Caroline Bruzelius
Anne M. Cogan Professor of Art and Art History
Carlo Tomasi
Professor of Computer Science