Enhancing Oral Health Sustainability through Deep Learning: A Focus on Tooth Decay Detection and Risk Analysis

Dental caries is one of the global health issues, necessitating advancements in diagnosis and treatment. Tooth decay may lead to the loss of the teeth or decay with severe pain. The preventive measures must be effective to rectify these problems. This paper used machine learning and deep learning to enhance dental caries detection along with other risk factors associated with it. A set of teeth images and demographic information of the teeth was used to make comparative analyses through various models. In teeth images, the pre-trained CNN using EfficientNet-B0 had better performance in results giving 92% accuracy, 94% precision, 93% recall, and an F1 score of 92%. Basic CNN presented only 87.5% accuracy along with a precision of 88%, the recall being 82% and the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$F 1$</tex> score was found to be 84%. By having the Random Forest Classifier able to predict risk from demographic dataset with an accuracy level of 96%, with a precision of 94%, recall of 90%, and an <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$F 1$</tex> score of 97% by surpassing the model by SVM, which received a result of 84%, precision of 79%, recall of 90%, and an F1 score of 88% in accuracy. This study innovatively integrates CNNs in the analysis of images and Random Forests in demographic risk assessment. Such advanced technology will help professionals in the dental profession diagnose the condition of a patient and assist them in developing an effective treatment plan. This raises the quality of services for the patients, and in the long term, it reduces expenses for healthcare providers.