An improved noninvasive method to detect Diabetes Mellitus using the Probabilistic Collaborative Representation based Classifier

Many government and health organizations around the world have arrived at the same conclusion, that the number of people suffering from Diabetes Mellitus is growing every year and will continue to grow. Recently, researchers have developed a novel noninvasive method to detect this disease by analyzing its facial block features. Compared with the traditional diagnostic tool that analyzes a blood sample, the novel method is noninvasive in nature. That being said, there are still many challenges yet to be resolved in this area. In this paper an improved noninvasive method to detect Diabetes Mellitus is proposed using the Probabilistic Collaborative Representation based Classifier with facial key block color features. The Probabilistic Collaborative Representation based Classifier was proposed in 2016 and is a combined classifier which joints the probabilistic theory and the Collaborative Representation based Classifier. The Collaborative Representation based Classifier was developed from the Sparse Representation based Classifier through the use of the l2-norm function to replace the l1-norm of the Sparse Representation based Classifier. Experimental results performed on a dataset consisting of 142 Healthy individuals and 284 diabetes patients showed that the proposed method outperforms seven other classifiers.