Gastric Cancer Diagnosis Using Terahertz Imaging

Terahertz imaging is a safe and inexpensive imaging technique with medical applications. However, presently there is no consensus on the methods which can be used to analyze the data resulted from this technique. This can constrain further developments in this technique. In this study, database of terahertz images, prepared from gastric cancer, is analyzed. There are 17 parameters for analyzing terahertz images. In this study, initially decision trees and principal component analysis (PCA) are used to identify important parameters, and then Support Vector Machines and Artificial Neural Networks are employed to classify terahertz data in order to find both healthy and unhealthy samples. Eventually, regarding the optimal results of this study, we can suggest new methods in terahertz reflective imaging for cancerous samples.