Formal analysis is a core method of visual description in art history. It seeks to explain the visual structure of an artwork: its visual elements, their relationships, and their function within the overall composition. As an analytical method, it strives to be exhaustive with a touch of scientific rigor, focusing on describing exactly what the viewer sees in an artwork, rather than discussing its historical context or emotional impact. There is currently no established data model for meaningfully describing the style, composition, and contents of artworks in a unified machine-readable format. Codifying formal analysis offers an avenue for designing this schema. Potential applications include automated comparison and classification of artworks, attribute-based object recognition, and the processing of such data into natural-language descriptions. It is this last point that is of particular interest for this year's conference. A common formal-analysis exercise in art history courses is to have one student describe an artwork only they can see to another student, who then attempts to replicate the artwork based on this description. There are thus some notable parallels between describing an artwork to a blind audience and inputting semantic data about an image into a computer. Both are limited by the information provided at the time, yet a human can extrapolate from past experience to fill in the missing aspects. Looking forward, we are not too far from a future where a computer could do the same, recognizing the contents of an artwork in detail and delivering a description that is suitable for providing a more accessible visitor experience. This presentation aims to explain formal analysis to a mixed audience, briefly discuss the potential applications of its codification, and present a review of previous and current work related to achieving that goal. Speaker(s) Session Leader : Illya Moskvin, Software Developer, Indianapolis Museum of Art MCN 2016 Presenting Sponsor: Piction New Orleans, LA