| dc.description.abstract | Caries is the most common disease and most experienced by people in the world. Caries is caused by excessive sugar consumption, lack of dental health care, and difficulty accessing standard dental health services. Caries is divided into various types, superficialis caries, media and profunda caries. Caries itself is usually more accurately detected through Chest X-Ray (CXR) image analysis. To overcome this problem, a method is needed to classify dental caries through CXR images automatically. The purpose of this study is to faciliate doctors and medical workers in classifying caries that are difficult to detect. The method proposed in this research is a Probabilistic Neural Network to caries superficialis, media and caries profunda. The stages before classification are preprocessing (scaling, grayscaling, contrast), segmentation (thresholding, canny detection) and feature extraction using Moment Invariants. This study used 372 data clinics in the city of Medan, consisting of 207 training data and 108 testing data and 57 validation data, which resulted in the final output consisting of 32 images of media caries, 56 profunda caries, 20 superficial caries. | en_US |
