Closer Looking: Computer Vision in Material Studies of Art

• Published in: The Art bulletin (New York, N.Y.) Vol. 106; no. 2; pp. 29 - 32
• Main Authors: Mintie, Katherine, Messier, Paul, Crockett, Damon
• Format: Journal Article
• Language: English
• Published: New York Routledge 02.04.2024; College Art Association, Inc
• Subjects: Art history; Computer vision; Datasets; Materials; Vision systems; Visual arts
• DOI: 10.1080/00043079.2024.2296276
• ISSN: 0004-3079

Studying the material nature of artworks has long been central to art historical study as a means of understanding how artworks are made, when they were produced, the aspirations of their makers, and their intended significance to viewers. Conventionally, art historians have gained knowledge of the physical qualities of artworks through close visual observation and, sometimes, through collaborations with conservators and conservation scientists. We propose that computer vision can serve as another powerful tool in this area. Though it might strike some as counterintuitive, we believe that computer vision based on abstracted data that model the material properties of artworks, in conjunction with subject matter expertise, can complement and deepen traditional art historical practices like close looking.