Presented at the Annual Biomedical Research Conference for Minority Students (ABRCMS)
Eli Stein; Northwestern University Feinberg School of Medicine
Dr. Bruce Tan; Northwestern University Feinberg School of Medicine
Dr. Claus Peter Richter; Northwestern University Feinberg School of Medicine
Artificial Intelligence and Robotics | Computational Biology | Diagnosis
Chronic Rhinosinusitis (CRS) is a nasal disease characterized by the inflammation of the mucosa and paranasal sinuses with a duration of at least 12 consecutive weeks. So, to diagnose CRS, one needs to keep a record of their symptoms for ~12 weeks before they are recommended to get a tomography which will allow physicians to classify them as a patient with CRS or without. This is a timely and costly process; thus, machine learning should be used to speed the process up. Since patients with CRS have more obstructed noses, the sound produced should be different than an individual without CRS. Thus, machine learning algorithms would be able to effectively classify patients as with or without CRS based on the way they speak. This project explores using Python and different machine learning algorithms to classify individuals as with or without CRS based on the way they speak. Each individual read a list of words. For the testing set, some of these individuals had CRS, while others did not, but their files were not labeled. For the training set, individuals read each word with their nose obstructed to model the sound produced by CRS patients and unobstructed to be a control. The voltage values from the recordings were then processed through a Fast Fourier transform, spectrogram, and continuous wavelet transform to convert the voltage values to the frequency domain. These frequency values were run through different machine learning algorithms (SVMs, neural networks, etc.). Each algorithm is then evaluated based on the time it took to run, as well as how accurate its predictions were.
Using Machine Learning to Recognize Chronic Rhinosinusitis.
Retrieved from: https://digitalcommons.imsa.edu/student_pr/101