Skin Condition Classificaiton by Automatic Image Machine Learning

In this paper, we proposed a novel Artificial intelligence (AI) based approach which could classify various images of human skin into appropriate categories based on associated features. The proposed system uses K-nearest neighbor classifiers, a neural network, Naïve Bayes (NB), and Gradient Decent / Ascent Classification to classify the features. In the experiment, the results show an improvement over existing approaches in accuracy. The experimental AI algorithm can achieve 50% accuracy with KNN, 60% with NN, 77% with naïve Bayes and 90% with Gradient Descent method.