Brian Frost
Projects
Ideal A-Scan Determination for Vibration Measurements within the Organ the Corti
This project shows the design and performance of an algorithm to be used in in vivo cochlear mechanics measurements with OCT. Specifically, given a volumetric scan of the gerbil cochlea taken through the round window, this algorithm uses an anatomically-informed image segmentation method to determine the best location from which to take vibration measurements. It served as the final project for a graduate course in Digital Image Processing at Columbia University under the instruction of Christine Hendon.
Hybrid Power System for a Cochlear Implant
This project is the design of a low power hybrid cochlear implant device, and includes the implementation of a prototype circuit. It was completed as the final project for a graduate course in Power Electronics at Columbia University under the instruction of Prof. Matthias Preindl.
Spectral Domain Optical Coherence Tomography Vibrometry is commonly used in our lab for cochlear mechanics measurements, but it is a very computationally intensive process. In this project, we use NVIDIA CUDA to leverage parallel processing techniques to achieve a 20x speedup in computation time.
Fluorescence Confocal Laser Scanning for in vivo Cochlear Imaging
While Optical Coherence Tomography is often used for cochlear mechanics measurements, it does not allow for the imaging of the Tectorial Membrane - a mechanically important gel-like structure above the Organ of Corti. This project seeks to design a confocal microscope to be used alongside an OCT system, so as to allow for imaging of the Tectorial Membrane in vivo.
Networks Impacted by Axonal Swelling
This is a comprehensive summary of computational neuroscience research efforts started in the Summer of 2018, under the advisement of Prof. Stan Mintchev. We have brought the project to the BAMM conference at Virginia Commonwealth University, and the associated poster can be found here. We hope to publish our findings in the near future.
R3CAP: Short Range Motion Capture of RFID Tags
This project is an innovation on current RFID tag tracking technology; we look to use FPGA beamforming techniques to localize multiple RF sources in 3D space, and track them in real-time. A low accuracy prototype can be seen here. This was a senior year capstone project completed in May 2019, under the advisement of Prof. Sam Keene and in conjunction with Nikola Janjusevic and Karol Wadolowski.
Gradient Sensing Via Cell Communication (arXiv:1806.07483)
This paper was written as a part of a mathematics REU at Oregon State University, under the advisement of Prof. Juan Restrepo.
Fully Integrated 100MHz Phase Locked Loop
This phase locked loop was designed as the final project for a junior course in integrated circuit engineering, under the advisement of Prof. Toby Cumberbatch and in conjunction with Nikola Janjusevic.
This serves as documentation for a junior year final project, worked on by four students including myself, Karol Wdolowski, Nikola Janjusevic and Jack Langner. A video demonstrating its operation can be found here.
Exploration of the FitzHugh-Nagumo Model
This document contains a pure approach to the FitzHugh-Nagumo neuron model using the language of geometric singular perturbation theory, as well as a computational discussion including a number of simulations. This constituted a final project for a course in numerical analysis instructed by Prof. Stan Mintchev.
A Numerical Study of the Central Limit Theorem
This is a published MATLAB script, which constituted a final project for a sophomore MATLAB seminar class. The topic was chosen entirely out of interest due to my having taken a probability theory course in the same semester.
This computer was developed as a final project for a course in computer architecture, and was completed in conjunction with Karol Wadolowski.