Akademik Veri Yönetim Sistemi
9th International Artificial Intelligence and Data Processing Symposium, IDAP 2025, Malatya, Türkiye, 6 - 07 Eylül 2025, (Tam Metin Bildiri)
This paper proposes an artificial intelligence-based image processing system for the early diagnosis of scoliosis, determination of its severity and monitoring the progression of the disease. The proposed system includes a YOLOv8 architecture-based object detection model built on the Roboflow platform and trained with spine X-ray images labeled according to the Cobb angle. The model detects vertebral alignments in the images and classifies the degree of scoliosis automatically. SCODIAC software, which is widely used in the healthcare field, is utilized in the Cobb angle measurement process. Data augmentation, early stopping and optimization techniques were applied during the training and testing process of the proposed model. The performance of the proposed model was obtained as mAP 82.5%, Precision 76.2% and Recall 83.2%. It is observed that the proposed model based on YOLOv8 architecture has a great performance. The obtained results show that the system can detect especially moderate and advanced scoliosis cases with high accuracy and can be integrated into clinical decision support systems. This study aims to contribute to digitalized health solutions by providing a fast, accurate and non-invasive approach to scoliosis diagnosis.