Facial Disease Progonstication using Deep Learning Technique

Due to debates on the correlation between illness and the face that go back thousands of years, facial diagnosis has become widespread. The objective of this study is to see whether deep learning techniques can identify diseases from unrestricted 2D face pictures. Introducing deep learning for facialskin related diseases have completed successfully computer-aided facial diagnosis for five distinct illnesses with relatively little available dataset. In reality, Getting pictures of face-skin illnesses is challenging, costly, and tedious, and it comes with ethical considerations because it entails managing personal data. As a result, the datasets used in face diagnostic research are frequently small and private in comparison to those used in other machine learning application sectors. A non-invasive, low-cost solution for disease screening may be made available by the success of deep-learning applications in face diagnostics on a small dataset. Comparing three deep learning technique to execute on the dataset to obtain better result from better technique on face-skin disease dataset. In the test dataset of images the best model achieves accuracy of 0.9752 with a loss of 0.19.