ABSTRACT
With advancements in imaging technologies, the efforts to digitize fine art paintings have increased dramatically since the turn of the century. These efforts have made hundreds of thousands of paintings digitally available. This unprecedented access enables researchers to analyze computationally a large number of paintings concurrently and in a way that was previously impossible. Traditionally, art historians have employed a range of tools to analyze paintings. Stylistic analysis requires knowledge of composition principles and techniques employed in different art movements or by artists in various time periods. The disparities and commonalities between styles or art movements are intricate, making it impractical for a single expert or a small group to master even a significant portion of the facts and relationships about all styles or movements. When styles can be automatically distinguished and categorized, the once tedious work of cataloging an extensive collection of paintings becomes manageable.
