A DEEP LEARNING BASED ENHANCED COMPUTATIONAL MODEL FOR DENTAL CARIES CLASSIFICATION

Dental Caries is one of the major oral diseases that can be seen increasing among adults as well as in children. Many Artificial Intelligence (AI) based works have been caried out for early detection of dental caries but achieving a good accuracy is still a challenge. This work aims to develop a model that can classify the three classes of the dental caries namely enamel caries, dentin caries and pulpitis. The proposed design is a fine-tuned model based on the VGG16 model emulating deep learning-based classification of dentin caries. The dataset of Radio Visio Graphy (RVG) images which comprises of infected tooth is collected and labelled for this purpose. The proposed model is also compared with the fully train VGG16 and Bi-Long Short-Term Memory (Bi-LSTM) coupled with the transfer learning. The performance of these models is evaluated based on the Accuracy, Precision, Recall and F1 Score. Based on the evaluation metrics, it has been observed that the proposed method is able to achieve highest accuracy as compared to the fully train VGG16 and Bi-LSTM coupled with the transfer learning. The proposed model shows an overall accuracy of 97.87% with minimal loss. It's performance has also been compared with state of the art models with similar settings to verify its performance upper hand.