MSCS Candidate Subrat Mishra successfully defends COVID-19-related thesis
On December 23, 2020, MSCS candidate Subrat Mishra (MET’21) defended his thesis titled “Analysis of SARS-COV-2 in Thoracic Computerized Tomographic Imaging Data with Machine Learning Algorithms.” Thesis committee members included Dr. Reza Rawassizadeh(Advisor), Dr. Lou Chitkushev, Dr. Kia Teymourian, and Dr. Guanglan Zhang.
One bottleneck in diagnosing COVID-19 (also known as SARS-Cov-2) has been a dependence on radiologists to read Computerized Tomographic (CT) test results. Mishra—as part of a team—devised an algorithm that could analyze thoracic CT imaging data to detect the presence of SARS-Cov-2. They used both traditional classification algorithms and VoxNet, a three-dimensional Convolutional Neural Network (CNN) classifier.
For the traditional classifiers, the team developed a novel method for extracting low-dimensional features from the pixel intensities in the CT images, while for the CNN classifier’s inputs, they transformed CT images using Fourier Transform to generate high-dimensional features with frequency components. They tested their approach on CT imaging data from 260 unique patients.
Congratulations to Subrat Mishra for successfully defending the thesis and research contribution!