Classification Method With Machine Learning For Images Of Oral Lesions

Abstract

Oral lesions as abnormal changes in the oral tissues, which can occur in various parts of the mouth such as the lips, tongue, gums, and others. These lesions can be early symptoms of oral cancer which often causes no symptoms in the early stages. To identify oral cancer, examinations are carried out through history taking, physical examination, and supporting examinations such as biopsy. However, this process takes time and risks slow treatment. With the development of technology, identification of oral lesions can be done quickly through oral image analysis using machine learning technology. Several previous studies have successfully used digital image processing methods to detect oral lesions with a high level of accuracy. The author then discusses recent research that uses the decision tree method in machine learning for classification and detection of malignant and benign oral lesions. The results showed a good level of accuracy, with a classification accuracy of 92% and a detection accuracy of 96% for malignant oral lesions and 86% for benign oral lesions.