I am a research fellow with the Department of Radiation Oncology at Mayo Clinic, currently working on a 4D X-ray imaging improvement project for image-guided proton therapy (IGPT). I develop clinically deployable software for motion-managed radiotherapy delivery, and conduct research on reliable tumor-tracking with computer vision / deep learning methods.
Previously, I studied nanoconfined polymer electrolytes for safer rechargeable lithium-ion batteries at Brookhaven National Laboratory (BNL), using synchrotron X-ray scattering experiments carried out at the Advanced Photon Source and the National Synchrotron Light Source II.
I enjoy solving scientific/quantitative problems with programming. In particular, I am fluent in Python, the SciPy stack and other associated scientific modules (Numpy, SciPy, Matplotlib, Lmfit, Mayavi, Scikit-image etc.). I like readable, elegant Numpy vectorization solutions, but am also comfortable with extending Python via Cython, Numba or pybind11/C++ for numerical performance boost. For larger coding projects, I use C#/.NET (LINQ is marvelous) and strive for robustness and maintainability through good software engineering principles.
CAMPEP Accredited Certificate Program in Medical Physics, 2018 - Present
Ph.D. in Mechanical Engineering, 2015
University of Colorado at Boulder
B.E. in Materials Science and Engineering, 2010
As an essential component of our image-guided proton therapy (IGPT) practice, I am developing a clinical computer vision application (Code name: WuKong) for visual guidance and tumor tracking during motion-managed treatments, based on high speed acquisition and real time processing of fluoroscopic frames. Being fully DICOM-aware, this .NET/C# Windows desktop application is integrated into existing treatment planning infrastructure for medical data transfer as well as HIPAA-compliant cloud storage, and provides a user-friendly graphical interface.
Correct delivery of proton treatment is predicated upon accurate spatial calibration. While the in-room position of imaging X-ray sources and flat-panel detectors are engineered to a given specification, their actual positions may deviate from the specification by small errors. This project delivers an iterative algorithmic calibration routine to correct such errors, i.e. translation, rotation and non-affine distortions. It may also be used as an independent QA test of vendor’s calibrations.
We fill solid polymer electrolytes (SPE) into rigid nanovolumes, and characterize their dynamics and structure using X-ray photon correlation spectroscopy (XPCS) and grazing incidence X-ray scattering (GIWAXS), respectively.