Space-Division-Based Stroke Classification and Localization Using Microwave Medical Sensing

This paper presents a novel methodology in the field of microwave medical sensing, which enables quick classification and localization of strokes without the need for imaging. The method introduced in this study utilizes space division within the region under examination, specifically the head, by leveraging the principles of the generalized scattering matrix theory and incorporating a prior knowledge of brain anatomy. Furthermore, a unique approach based on decision tree learning method is proposed to enhance the identification of stroke features for effective classification. The simulation results demonstrate noteworthy enhancements in both the accuracy of stroke classification and the efficiency of localization, surpassing the performance of the traditional method. The classification results have a 14% performance improvement, whereas the localization accuracy of the proposed method is 94%.