Default header image

Neonatal Cerebral Lateral Ventricle Segmentation from 3D Ultrasound Images

Main Contributor: Zachary Szentimrey

Our work is to automate neonatal ventricle segmentation from 3D ultrasound images using deep learning segmentation models. We use both labelled and unlabelled data to assist in 3D ultrasound segmentation. The purpose of our work is to provide clinicians with quantitative information on ventricle size changes in neonates with intraventricular bleeding to improve patient outcome.