I am a passionate engineer and scientist who is developing machine learning and data analytic tools for use in a variety of fields including mycology, activism, and scientific research. In March of 2019, I finished my Ph.D in Biomedical Engineering at Rensselaer Polytechnic Institute advised by Dr. Ge Wang.
My thesis focused on software tools for use with state-of-the-art, photon-counting X-ray detectors for nondestructive imaging of biologics. Several years of my research career were devoted to the (nearly non-existent) field of radiation electrophysiology in search of a protein mechanism responsible for the sensation of (and/or measured ionic responses to) photons with high energy (e.g. X-ray).
Outside of the lab, I am an avid (albeit amateur) forager, mycologist and fungi photographer.
Developed computer vision algorithms for identifying Lanthanide-based nanomaterials in an X-ray phonton-counting CT system. Both experimentation and simulation provided data for statistical methods such as Fisher's linear discriminant analysis and machine learning algorithms.
Built an electrophysiological rig coupled to an X-ray source as a tool to measure single-cell radiation responses to bursts of radiation (1-120kev). Specific shielding and motor controls were also designed to generate the pulses relatively passively from the perspective of the cells until the time of the pulse.
Used optical character recognition (OCR), word ranking, and graph theory to read, organize, and group my PDF library for the systematic review of literature that I had compiled over the years of my thesis work.
This thesis provides solid results and physical insights for improving the current photon-counting imaging performance provided by this MARS system through (1) pre-reconstruction, spatial non-uniformity corrections, (2) hardware mechanisms to increase spatial resolution, and (3) post-reconstruction methods for analyzing image regions which contain varying levels of contrast materials. During this research, a full-scale system simulator was developed in MATLAB as the “ideal” photon-counting/system model, which is a significant portion of the thesis. Our major goal is to provide a base for future development of a seamless imaging pipeline, which is from acquisition and preprocessing to reconstruction and post-processing, analysis, as well as material decomposition in a unified deep learning framework.
Read Thesis