Regresar

Image Segmentation Techniques Application for the Diagnosis of Dental Caries

Abstract:

The diagnosis of a patient with caries is a process performed by oral health professionals, who after auscultation on the dental surfaces, determine the degree of affectation, following the visual inspection protocols, which present international standards. In recent years, there has been a growing interest in developing new techniques that allow the establishment of medical diagnoses, supported by information technologies, specifically in the early detection of diseases to apply the respective treatments. This work presents a method to determine the level of affectation of caries in the oral cavity, which applies a classifier that consists of 5 phases: capture, preprocessing, segmentation, extraction of characteristics and classification of objects. The proposed methodology considers a bank of images of dental pieces, all extracted from private dental clinics, as well as from the Integral Clinic CIAM II, which pertains to the Odontology Pilot School of the University of Guayaquil. For the classification, a multilayer perceptron artificial neural network was used, while for the validation of the work, 2030 images were analyzed, finding 80% success in the results, which were corroborated following the norm of caries classification and the criteria exposed by experts.