A Review of Machine Learning Methodologies for Dental Disease Detection

Abstract

Dental diseases have become commonplace in today‘s fast paced world. Currently, most medical practitioners rely on manual analysis of a patient's oral cavity for initial diagnosis. Later, they rely on manual analysis of radiographs or x-rays for advanced diagnosis. To reduce this effort, systems are proposed for disease detection techniques working with radiographs or x-rays, which are only accessible to dental practitioners. Other techniques that work on raw, visible light based images of oral cavity have been trained on miniscule datasets with a narrow list of diseases that can be detected. There have been efforts in recent times to repurpose the general use machine learning algorithms such as CNNs for the particular task of disease detection and classification in medical imaging. The field of dentistry can benefit greatly by focusing more research on visible light images, allowing practitioners to offload the initial review of a patient to machines, giving them more bandwidth to work with cases that require more of their attention. This review intends to provide a comprehensive survey of currently proposed machine learning based dental disease detection systems along with suggestions towards what can be improved in the future to provide a better insight to researchers working in this domain.