Eureka! My research interests (current, at least).
1). Image Segmentation Utilizing Deep Learning – Image segmentation, and thinking about computers can think like people to discriminate structures in images, has always fascinated me. Utilizing deep learning, we can achieve this and I am currently working on projects that take advantage of upcoming algorithms for such tasks in analyzing cardiovascular CT images.
2). Mesh Generation from CT Images – My research interests build off each other. Here, I am interested in validating and improving several deep learning algorithms to generate mesh models of the aorta and other relevant vasculature used in TAVR mechanical simulations.
3). Improving TAVR Simulation Workflows – To improve computational simulation (such as FEA or CFD or FSI) workflows for TAVR studies, we need to quicken the time it takes to generate the patient-specific mesh environments that serve as boundary conditions. To do this currently, we use manual segmentation of CT images, which takes several hours to do amongst experienced graduate students. But, with advances in deep learning, we can get this time to minutes, even seconds, thus allowing for more simulations to be done in the same amount of time it takes to do one. With quicker simulations, we have the potential to create in silico trials for TAVR devices and, perhaps, create a software that clinicians can use to observe potential TAVR devices in different flow conditions in patient specific anatomy to best select the device for them.