Advances in Computational Art Connoisseurship: Digital image processing of manufactured patterns in art supports

Basic digital image processing is an underexploited computational tool in the incorporation of digital technology into art history research. This thesis covers a pioneering effort undertaken since 2007 addressing the closure of this gap. The path taken was the development of digital image processing methods to identify matching patterns in art supports and apply the data visualizations produced to art historical connoisseurial studies of fifteenth- to nineteenth-century European paintings on canvas and to fifteenth- to seventeenth-century European prints, drawings, and manuscripts on laid paper. The software developed produces striped maps of canvas thread density that have been used to identify canvas from the same roll, principally for the paintings of Vincent van Gogh and Johannes Vermeer, and watermark overlays that confirm exact matches indicative of sheets of paper made on the same mold, principally among the prints of Rembrandt van Rijn, the codices of Leonardo da Vinci, and seventeenth-century Dutch drawings. These new digital tools have provided significant insights into the attribution and dating of paintings and the dating of drawings.