In 2015 Gays et al. found a new use for the Convolutional Neural Networkand that was to generate art. More specifically, the researchers came up with a way to superimpose the artistic style of one image over the content of another image. This process is known as Neural Style Transfer. Since then many researchers have made improvements on this method in attempts to produce better quality images and reduce processing time. One notable work is that of Eddie Huang and Sahil Gupta in which they perform NST by interpreting style as a distribution of features and defining the constraint using the Wasserstein metric. They argue that their method is superior to contemporary approaches in terms of style extraction. In this paper I create a metric to measure style retention quantitatively, and attempt to compare the work of Huang and Gupta to other researchers to test the claim that Huang and Gupta’s process is more effective at capturing style than others. 1 Visually inspecting the generated stylized images would suggest the claim is valid, however the results of the regression suggest there is not a significant difference between the style retention capabilities of the different methods.