TEXT

Abstract

A variety of machine learning methods have been used for art classification. These methods have predominately focused on Western art and neglected the incredibly rich and diverse field of Chinese art. Determining what dynasty a painting is from is one of the foremost tasks that art historians undertake to study Chinese paintings, but it can be difficult to pinpoint which dynasty a work is from. The goal of our research is to evaluate how well existing machine learning methods using deep learning and hand-crafted features can classify Chinese paintings based on dynasty. In our experiments, we aim to find the best-performing model from these methods. This will allow art historians and viewers to study Chinese paintings more efficiently by establishing a baseline for placing paintings in their art historical context.