ABSTRACT
Deep learning can be applied in a wide number of areas, such as speech processing, bioinformatics, astronomy, astroinformatics, disease prediction, image analysis and drug discovery. The huge growth of data and the availability of cheap hardware with GPUs and TPUs are among the reasons why deep learning is becoming more acceptable. Methods for the prediction of dental diseases have been confusing and, to a certain level, unsuccessful. Experiments were conducted using YOLO. YOLO cleverly does object detection in real time using Convolution Neural Networks. It is found from literature that YOLO provides very high real-time accuracy. In this study, we have tried to automate the system of identifying dental diseases such as tooth decay, fillings, fluorosis, chipped tooth and oral thrush through an app with an accuracy of 90%, which is shown through good specificity and sensitivity parameters. The challenge faced by most experiments in the medical or dental field is the availability of appropriate data sets. Labelled data consisting of images of the five different classes were used to train the model. Experimental results are provided for all the above diseases and the overall accuracy is very encouraging.
