Physics-grounded AI for reliable medical imaging
I develop machine learning methods for medical imaging that explicitly model acquisition physics, anatomy, and data consistency. My goal is to build algorithms that remain reliable across scanners, protocols, and patient populations, and to lay the foundations for a broader research program in generative computational imaging.
A research program for reliable medical imaging
I work at the intersection of machine learning, optimization, and medical imaging. Across my work, I ask how we can move beyond purely data-driven reconstruction toward methods that incorporate imaging physics, leverage structure in anatomy and acquisition, and remain dependable in real clinical settings.
Reliability across clinical variation
Design imaging algorithms that continue to perform across heterogeneous scanners, acquisition protocols, and patient populations.
Physics-grounded inference
Develop inverse-problem and generative methods that enforce agreement with measured data instead of depending only on paired training sets.
MRI and computational imaging
Focus on MRI as a core domain for reconstruction, super-resolution, and generative modeling, with attention to the practical constraints of clinical imaging.
Generative computational imaging
Build a broader program that co-designs imaging systems and inference algorithms for robust, efficient, and accessible medical imaging.
Three research thrusts
My work centers on combining imaging physics with modern generative modeling to create methods that are not only accurate, but also structured, controllable, and clinically reliable.
Multi-image and multi-modal computational imaging
Develop structured generative frameworks that integrate scans across visits, contrasts, sites, and modalities for reconstruction, longitudinal analysis, and downstream clinical tasks.
Co-design of acquisition and reconstruction
Couple acquisition design with reconstruction through differentiable physics and learned priors to reduce scan time while improving image quality and access.
Generative priors for medicine
Build generative models that are anatomically plausible, controllable, pathology-aware, privacy-preserving, and aligned with the requirements of clinical deployment.
Representative projects and technical contributions
These projects show how my research combines physical modeling, generative methods, and clinically motivated problem design.
Zero-shot and operator-aware through-plane super-resolution
Developed internally trained and cycle-consistent methods for anisotropic MRI super-resolution that use acquisition operators directly, reducing dependence on paired high-resolution training data.
Generative priors with data-consistent inference
Developed diffusion-based formulations for MRI inverse problems that use generative models as priors while explicitly enforcing consistency with measured observations.
Large-scale data curation for generative imaging
Curated and analyzed large-scale 3D brain MRI datasets for diffusion-model development, with emphasis on anatomical plausibility and downstream utility for reconstruction tasks.
Recent honors, invited talks, and service
IPMI 2025 Best Poster Award
Awarded for Cycle-Consistent Zero-Shot Through-Plane Super-Resolution for Anisotropic Head MRI.
Invited talks in Japan
Presented invited talks on through-plane MRI super-resolution at the University of Tokyo and the Nara Institute of Science and Technology.
SASHIMI organizing committee
Serving on the organizing committee for SASHIMI at MICCAI across three consecutive years.
SynthOCT organizing committee
Serving on the organizing committee for the SynthOCT 2026 challenge at MICCAI.
NSF Graduate Research Fellowship
Supported independent research in AI for medical imaging through the National Science Foundation Graduate Research Fellowship Program.
Teaching and mentorship
I teach students to connect imaging physics, algorithms, and modern AI, with emphasis on conceptual understanding, technical rigor, and critical evaluation of model behavior.
Professional service
My service includes workshop organization, peer review for leading venues, and sustained engagement with the medical imaging research community.