AI for Medical Imaging: $400K Award to Georgia State Researcher
Dong Hye Ye will build AI tools that read medical images alongside patient history, aiming to automate radiology reports and improve diagnostics.
By Sarah Ross ·
Artificial intelligence systems will soon be tested on their ability to read medical images the way human doctors do, by combining visual data with patient history. A new five-year project at Georgia State University seeks to close the gap between current AI models and clinical practice, where context is essential for accurate diagnosis.
Associate Professor Dong Hye Ye has received the National Science Faculty Early Career Development Award. The grant provides more than $400,000 to fund his research. The project, titled "Attending & Bridging Context in Vision-Language Modeling for Medical Image Analysis," focuses on creating AI that interprets scans alongside clinical text and patient information.
The research targets three specific medical areas: radiology reporting, eye disease, and adolescent brain development. The goal is to produce AI results that are not only accurate but also reliable and interpretable. A key application is the automatic generation of radiology reports, a task that currently requires significant human time and expertise.
Ye explained that clinicians never view an image in isolation. They cross-reference it with the patient's history, prior scans, and clinical notes. "This project is about building AI that does the same," Ye said. He noted that the system must remain reliable even when that supporting information is incomplete or noisy, a common challenge in real-world clinic settings.
Ye joined the Georgia State Department of Computer Science in 2022. He is a core faculty member of the Center for Translational Research in Neuroimaging and Data Science, a tri-institutional research hub. His previous work focused on advancing medical imaging through machine learning, with applications in neuroimaging and CT image reconstruction. He completed his doctoral research at the University of Pennsylvania.
Todd N. Rosenstiel, dean of the College of Arts & Sciences, praised the award as a reflection of the college's commitment to public impact. "By advancing AI methods that can help clinicians interpret medical images with greater accuracy and reliability, Dr. Ye is directly contributing to research with meaningful potential to improve patient care," Rosenstiel said. The award supports Ye's work as he integrates innovative research with educational excellence.
