Camera-Based Machine Learning for Skin Disease Detection: An On-Device Diagnostic Approach

Abstract

One of the biggest problems in the Indian medical system is the lack of healthcare workers. The ever-increasing population of India has exerted significant pressure not just on the medical infrastructure but has also overwhelmed healthcare workers. This situation has caused the wait times for patients to increase significantly; speeding up this process requires the use of artificial intelligence, especially on the diagnostic side. Dermatology is one branch of medical science where diagnosis can be aided with camera image-based machine learning. Diagnosis of surface-level skin diseases that are easily visible to the naked eye can be made much quicker using image-based machine learning. It will enable doctors to expedite their diagnosis and also reduce the likelihood of an incorrect diagnosis. For the purpose of this paper, such an app called Dermascan has been created to test the real-world applications of this hypothesis.