Mapping of Peripheral Pulse Pattern Using Artificial Intelligence in Lung Cancer and Tuberculosis

Ancient India emphasised on radial pulse dynamics (nadi pariksha) for medical diagnosis in the patients. Modern medicine provided many instruments for recording radial pulse from the subject and analyse its morphology. The pulse morphology is seen to vary from person to person and also with respect to time in the same individuals. Morphology index introduced in the past was observed to vary from 1 to 0.25 depending upon health status of the individual. With an objective to link pulse pattern with the disease, previously developed application software based on deep learning Convolutional Neural Network (CNN) models has been used to obtain profile of fifteen pulse patterns in 165 data files recorded from healthy volunteers and patients with lung disorders. It is observed that in each data file one pattern is dominant followed with less dominant patterns. Dominant patterns of healthy and diseased subjects are observed to be specific in this study.