A Review on Non-invasive Tools and Deep Learning/Machine Learning Methods for the Early Identification of Oral Cancer

Abstract

The most common type of cancer in India is called "oral cancer." Poor lifestyle choices like chewing tobacco, binge drinking, smoking cigars or pipes, getting an HPV infection, and being in the sun can cause irreversible cell proliferation and the formation of cancer. Oral cavity cancers are becoming more common and need to be treated right away. While a visual examination by a specialist and a gold standard "biopsy" carried out under expert supervision and using the detection instrument can detect an oral cancer lesion, this method is not ideal because it is invasive. There is a summary of various invasive and non-invasive tools, and there is a review of machine and deep learning techniques used by numerous researchers. It was found that some models with pre-defined networks and transfer learning techniques were used on a small dataset during the review process. However, there is an urgent need for non-invasive early diagnosis of multiple types of lesions in the early stages of oral cancer.