Innovative Skin Disease Prediction and Customized Care Recommendation Leveraging Capsule Network Enhanced Through Transfer Learning

A novel type of network that maintains spatial information is Capsule network (CapsNet). This project aims to develop an advanced skin disease prediction and personalized care recommendation which uses Capsule network that is enhanced by transfer learning. One of the limitations of the classical convolution model is it fails to capture spatial hierarchies is addressed by using Capsule network which increases the diagnostic accuracy. Transfer learning allows to make use of pre-trained models and enhance generalization across various many skin condition and can reduce the requirement for expanded set of data. The system is trained on a vast dataset that will detect various skin disorders. Onto diagnosis, it generates a descriptive explanation on what skin disorder one might have, recommends nutritional intervention through food that might help in aiding for the betterment of the disorder, medication and treatment that one can take for the betterment of the skin condition and a mini chatbot that enhances the user experience by providing assistance and answer the questions prompted by the users. This system supports dermatologists and extent the access to excellent dermatological care including undeserved area.