Prediction of Diseases Based on Facial Diagnosis Using Deep Learning

Abstract

Practitioners can diagnose diseases by looking at a patient's facial features, a process known as facial diagnosis. Doctors need to have a lot of real-world experience to diagnose facial conditions with high accuracy. Modern medical studies show that many diseases do, in fact, manifest matching unique traits on human faces. Due to the scarcity of medical resources, it is challenging to get a check-up today in many rural and undeveloped areas, which frequently causes treatment to be delayed. Limitations still exist, such as high expenses, lengthy hospital waiting-times, and conflicts between doctor and consulted patient that result in medical disputes, even in major cities. We can rapidly and easily do non-invasive screening and disease detection thanks to computer-aided face diagnosis. To perform CA facial-diagnosis of various diseases and test DTL methods for both single and multiple disease spotting on a short data-set, we therefore propose adopting deep transfer learning from face recognition.