Elevated intracranial pressure (ICP) is a critical finding in conditions such as traumatic brain injury and spaceflight-associated neuro-ocular syndrome (SANS). Optic nerve sheath diameter (ONSD) measured on ultrasound is a noninvasive biomarker of ICP, but current manual measurements vary between observers. This project developed machine learning models to automatically segment optic nerve structures and measure internal and external ONSD. Using expert-annotated ultrasound images, multiple deep learning architectures were trained and evaluated for segmentation accuracy. Postprocessing steps were applied to standardize measurements of ONSD 3 mm posterior to the retina. This approach showed strong agreement with expert annotations and demonstrates the potential for reducing variability in ONSD measurement. Future work will expand the dataset, refine segmentation in challenging cases, and adapt the method for broader clinical and spaceflight applications.