Computational Connoisseurship: Enhanced Examination Using Automated Image Analysis. (3rd April 2019)

• Record Type: Journal Article
• Authors: Ellis, Margaret Holben; Johnson, C. Richard
• Is Part Of: Visual resources. Volume 35:Number 1/2(2019)
• Page Start: 125
• Page End: 140
• Publication Date: 2019-04-03
• Subjects: Image Processing -- Canvas Thread Count -- Photographic Paper -- Decision Tree -- Connoisseurship; Art -- Documentation -- Periodicals; Art -- Audio-visual aids -- Periodicals; 025.177
• DOI: 10.1080/01973762.2019.1556886
• Languages: English
• ISSNs: 0197-3762

Abstract

Abstract : The feasibility of the application of image/signal processing for measuring, marking, matching, and sorting vast quantities of data derived from materials typically found in artworks is presented through four case studies. Different patterns produced by canvas weave structures, surface textures of historic photographic papers, chain line intervals in Rembrandt's printing papers, and watermark variations have been subjected to different modes of computational analysis. The art-historical implications that result from computer-generated algorithms – including dating, attribution, authenticity, and workshop practices – can be considered as "computational connoisseurship." The case studies discussed point to future areas for research. Finally, because of the need for statistically meaningful datasets of images, a practical means of recording internal paper structure is introduced.