This project aims to democratize MRI research by developing Cloud MR, an open-source virtual scanner platform accessible via web browser, enabling MRI experimentation, education, and development without the need for physical MRI machines.
This project develops a three-pronged strategy—combining global inverse modeling, physics-informed neural networks, and vision transformers—to generate high-resolution maps of tissue electrical properties using MRI, with broad applications in medical imaging and treatment.
This project develops open-source, memory-efficient electromagnetic simulation tools using tensor decompositions and optimization techniques to improve RF coil modeling for ultra-high field MRI, enhancing image quality and safety while reducing computational and licensing barriers.