Color Harmony Classification using Machine Learning Algorithms: KNN and SVM

Abstract

Color harmony captured the attention of many researchers who study images aesthetic concepts and evaluation, in this paper, a color harmony classification was accomplished using two well-known machine learning classifiers k-nearest neighbor (KNN) and support vector machine (SVM), the color harmony was determined using the basic concepts of color harmony scheme which are: Color Complementary harmony, Color Analogous harmony, Color Triad harmony, Color Split-Complementary harmony, Color Rectangle (tetradic) harmony, and Color Square harmony, which discussed in the paper with more details. A performance comparison between the two-classifier was discussed as well. The results show that the KNN performance exceeded the SVM performance for many reasons. Where the KNN success rate was 100% and on the other hand SVM was 85.71%.