Development a computer-aided diagnosis system for dental caries detection applying radiographic images.

Delayed diagnosis of certain diseases like dental caries frequently leads to intensified problems due to postponed treatment. Untreated cavities can progress and affect a deeper layers of the tooth, which result in discomfort, infection, and ultimately tooth loss. Presently, manual diagnosis of dental decay in radiographic images by dentists in most dental clinics is prone to errors due to factors such as fatigue and information overload, emphasizing the necessity for automated diagnostic tools. This study develops an intelligent computer-aided diagnosis (CAD) system for accurate and efficient dental caries detection in radiographic images. By integrating advanced image processing and machine learning techniques, the system effectively analyzed radiographic images to distinguish between healthy and decayed teeth. After gathering the dental image dataset, preprocessing techniques are applied to eliminate image artifacts and enhance image quality. The segmentation stage employs the Sauvola thresholding method to delineate tooth boundaries and the Canny edge detection method to precisely identify tooth contours. Subsequently, features of healthy and decayed teeth are extracted utilizing the texture and chaotic indices. Furthermore, between the various metaheuristic optimization techniques, the non-dominated sorting genetic algorithm (NSGA II) is selected for eliminating insignificant features. The performance is rigorously assessed using a 10-fold cross-validation strategy. Among various classifiers evaluated, the Support Vector Machine (SVM) achieved the highest accuracy of 93.33 %, demonstrating superior performance in caries detection. This outcome highlights the system's robustness and precision, positioning it as a talented tool for enhancing diagnostic reliability and making it a valuable asset for modern clinical practice.