Each person's eyes are different from each other's even identical twins and
have complex structures. In the field of biometrics, iris brings very important
information relating to health and other information. The purpose of this study was to
detect the possibility of health problems experienced by patients based on the
recognition of the iris. Digestive disorders, cholesterol, and stress are case studies of
research taken. The iris recognition method uses a deep learning algorithm approach
with a convolutional neural network architecture. The technique will be used to
recognize the iris, then classified based on the labeling that has been done and
confirmed the accuracy by doctors who are experienced in iridology. The use of deep
learning techniques in this study provides the highest level of accuracy of 81.92%.