3-D MRI-informed AI for Hand Bone 2-D X-ray Image Enhancement

Successful candidates will be enrolled as PhD candidates at the College of Engineering & Computer Science (CECS), VinUni.  

Project Information  

X-ray imaging is the most widely used diagnostic technique, yet in many cases fractures remain difficult to detect due to overlapping anatomical structures (e.g., skull–face regions) or because certain fracture lines cannot be captured clearly from available imaging angles, such as scaphoid fractures in the wrist or small bone fractures in the foot. These limitations may lead to delayed diagnosis, the need for additional MRI scans, or late clinical intervention.

This project proposes the development of a deep learning system capable of learning from a limited number of matched X-ray–MRI pairs to reconstruct hidden diagnostic information that conventional X-rays cannot directly reveal. The aim is to improve diagnostic reliability for bone fractures using standard X-ray images alone.

The core of the project is to develop a next-generation deep learning model based on a transformer architecture, in which the attention mechanism is optimized to better extract and utilize essential imaging features from X-ray and MRI data.

Project members include: Prof. Saeid Sanei (Principal Investigator) | Tran Trung Dung, Prof., MD, PhD (Orthopedic Surgery) | Bui Van Giang, MD, PhD (Diagnostic Imaging) | Nidal Kamel, PhD | Dr. Doan Dang Khoa | Dr. Pham Huy Hieu.

Research Objectives  

The project focuses on building an AI system capable of “seeing” hidden diagnostic information in X-ray images by learning from MRI data. The objectives are:

(i) To enhance the quality of both X-ray and MRI images using deep neural network–based super-resolution techniques.

(ii) To develop a 2D X-ray to 3D MRI mapping model based on a transformer architecture with optimized attention mechanisms.

(iii) To test and clinically validate the model’s performance in detecting difficult-to-observe fractures, particularly scaphoid fractures.

(iv) To conduct a comprehensive statistical evaluation—including dataset size, number of subjects, age, gender, imaging sessions, and other factors—to verify the model’s applicability in real clinical settings.

The project will be carried out through close collaboration between researchers, orthopedic surgeons, and diagnostic imaging specialists to ensure accuracy and clinical relevance.

Benefits & Impact  

Hand bone fractures, especially those involving the carpal bones, are highly common, with approximately 190 cases per 100,000 people each year. However, more than 15% of cases are missed in clinical practice due to the limited diagnostic information provided by standard X-ray images.

The outcomes of this project may significantly reduce the need for MRI scans at healthcare facilities lacking advanced imaging equipment, shorten diagnostic time, and improve early detection of fractures, ultimately contributing to better treatment outcomes.

This research direction holds substantial clinical and societal significance, particularly for frontline hospitals and regions with limited diagnostic imaging resources.

Project Contact 

For details, please contact Prof. Saeid Sanei via email [email protected]

Graduate Admissions Contact 

• VinUni Graduate Admissions
•  0978 549 846
• [email protected]

Learn more about Master’s, PhD programs, and scholarships at: VinUni Graduate Research Excellence Program

#VinUni #VinUniversity#VinUniGraduateResearchExcellenceProgram #VinUniPhD#VinUniMasters#VinUniResearchProjects