Paper Title
Artificial Intelligence for Dental Caries Detection and Segmentation on Periapical Radiographs

Abstract
Since the rapid development of neural network technology, image recognition has been increasingly applied in the medical field, significantly contributing in filed such as tumor detection in radiology, skin cancer screening in dermatology, retinal disease diagnosis in ophthalmology, and dental caries detection in dentistry. These technologies assist clinicians in analyzing lesions more accurately and efficiently, thus improving diagnostic precision and clinical workflow. This study proposes an optimized neural network-based approach for detecting dental caries in periapical (PA) radiographs, aiming to assist dentists in accurately identifying carious regions using polygonal annotations. The proposed method integrates advanced image preprocessing and augmentation techniques to enhance the clarity and quality of periapical radiographs while preserving essential anatomical structures. These techniques include rotation, brightness adjustment through gamma correction, linear transformation, and image resizing to increase data variability and improve model generalization. Additionally, two filtering methods — Contrast-Limited Adaptive Histogram Equalization (CLAHE) and Bilateral Filtering (BF) — are applied to enhance image contrast and reduce noise while maintaining important edge details, thereby optimizing the radiograph quality for more accurate dental caries detection. The detection framework is based on a fine-tuned YOLOv8 model with optimized parameters to enhance performance. Data from 300 dental patients were used for model training and testing. The proposed model was evaluated against state-of-the-art methods such as YOLOv8 and U-Net for tooth detection and caries segmentation. Experimental results demonstrate improved detection accuracy and robustness under real clinical conditions, highlighting the potential applicability of the proposed approach in practical dental diagnostics. Keywords - Dental Caries, YOLOv8, Periapical Radiographs, Image Preprocessing, CLAHE, Bilateral Filtering, Deep Learning.