A Hybrid Approach for Dental Diagnosis: Dental Diagnosis System (DDS) Based on Segmentation, Classification, and Decision Making

Özet

Computerized medical diagnosis systems utilizing X-ray images are crucial for accurate decision-making in disease identification and treatment. Subclinical diseases often lack recognizable clinical findings, making it essential to segment dental X-ray images into distinct groups. This study proposes a novel framework, the Dental Diagnosis System (DDS), which employs a hybrid approach combining segmentation, classification, and decision-making techniques. The DDS utilizes a state-of-the-art dental image segmentation method based on semi-supervised fuzzy clustering for accurate segmentation. Additionally, a new graph-based clustering algorithm, APC+, is introduced for classification. Finally, a decision-making procedure is designed to identify the final disease from segmented groups. The DDS is evaluated using a dataset from Hanoi Medical University, Vietnam, consisting of 87 dental images encompassing five common diseases: root fracture, incluse teeth, decay, missing teeth, and resorption of periodontal bone. The results demonstrate the DDS's superior accuracy of 93.24% compared to other methods, including fuzzy inference system (90.07%), fuzzy k-nearest neighbor (81.25%), prim spanning tree (57.26%), Kruskal spanning tree (57.56%), and affinity propagation clustering (97.08%). In conclusion, the empirical findings confirm the DDS's outstanding performance compared to related methods. The results of this study have the potential to significantly assist dental clinicians in their professional work.