ABSTRACT

Recent advances in computer vision and image processing have enabled basic automatic analysis of visual art. The author uses computer analysis to extract thousands of low-level numerical image content descriptors from digitized paintings, which are then used to compare objective levels of similarity between the artistic styles of different painters. The analysis reveals that Vincent Van Gogh's and Jackson Pollock's artistic styles are far more similar to each other in terms of low-level image features than Pollock's work is to that of other painters'. This report also proposes that this methodology is useful in quantifying similarities between painters or artistic styles based on large sets of numerical image content descriptors and for detecting influential links not easily detected by the unaided eye.

This content is only available as a PDF.