Klasifikasi Penyakit Pulpitis Pada Citra Radiografi Periapikal Menggunakan Metode Convolutional Neural Network (CNN)

Abstract

Teeth are an important part of the human body. Lack of dental care can lead to various dental diseases, one of which is pulpitis. Pulpitis is an inflammation of the dental pulp (the innermost part of the tooth that contains nerves and blood vessels) and the tissue around the tooth root. It can also be caused by toothache or tooth loss, especially in young people. To diagnose pulpitis, dentists use the periapical radiography technique. This technique provides clear images of all layers of the tooth, allowing diagnosis of the condition of the tooth and surrounding tissues. However, these radiographs can only be interpreted by dental radiology specialists, who are limited in number. Therefore, to facilitate the detection of pulpitis, classification using image processing techniques is used to assist doctors in classifying pulpitis based on radiographic images. This study uses the Convolutional Neural Network (CNN) method to classify pulpitis disease based on radiographic images. CNN is a variation of Multi Layer Perceptron (MLP) that has few free parameters because it does not require pre-processing, segmentation, or feature extraction. This study used 1000 image data divided into two classes: pulpitis and normal. The test results show that hyperparameters such as epoch value and optimizer greatly affect the accuracy. The highest accuracy achieved was 98.75% using the RMSPROP optimizer and an epoch value of 50. This research shows that the use of Convolutional Neural Network (CNN) can help dentists in diagnosing pulpitis. This system can be used to simplify and assist doctors in determining the diagnosis of pulpitis based on radiographic images