Utilization of YOLO for Detection and Segmentation of Ultrasound Medical Images in Hydatidiform Mole (Molar Pregnancy)

Diva Nabilla Balqis (Universitas Dinamika Bangsa)
Dwi Setia Ningrum (Department of Informatics, Dinamika Bangsa University, Jambi, Indonesia)
Dwi Erna Septi Utami (Informatics, Dinamika Bangsa University, Indonesia)

Media Journal of General Computer Science (MJGCS)
Vol. 3 No. 2 (2026): MJGCS

This study aims to explore the utilization of the YOLO (You Only Look Once) algorithm for detecting and segmenting ultrasound (USG) medical images in cases of hydatidiform mole (molar pregnancy). Hydatidiform mole is a pathological condition that requires rapid and accurate diagnosis to prevent serious complications. By using a diverse dataset of ultrasound images, the YOLO model was trained to recognize and identify suspicious areas within the images. Evaluation results indicate that this model significantly improves detection and segmentation accuracy compared to manual analysis methods, while also providing higher efficiency in the diagnostic process. The implementation of a YOLO-based system is expected to assist healthcare professionals in making quicker and more accurate decisions, thereby enhancing health outcomes for patients. This research highlights the potential of artificial intelligence technology in the medical field, particularly in diagnosing complex obstetric conditions.