Publication Date
Spring 5-2024
School
School of Business; School of Visual and Performing Arts
Major
Computer Science
Recommended Citation
Bryant, David, "Machine Learning Image Classifier for Identifying Composition in Landscape Paintings" (2024). Senior Honors Theses. 1394.
https://digitalcommons.liberty.edu/honors/1394
Abstract
The purpose of this thesis is to examine the usefulness of machine learning in the art world with a simple model. This model is an image classifier designed to identify the type of classical composition present in landscape paintings. After several tests and various modifications, the image classifier achieved around 50% accuracy on images within its dataset. For images outside of the dataset the model’s results were 65% accuracy on an easier image and 15% accuracy on a more difficult image. This renders the model impractical for educational purposes and raises questions as to how current machine learning techniques in the art world could be improved by more accurate aesthetic evaluation.
