Artificial Intelligence Algorithms for Medical Imaging and Healthcare

Author

Jonathan William Stubblefield, Arkansas State UniversityFollow

Date of Award

11-12-2021

Document Type

Dissertation

Degree Name

Molecular Biosciences, Ph.D.

First Advisor

Xiuzhen Huang

Committee Members

Cody Ashby; Emily Bellis; Jake Qualls; Jason Causey; Lynita Cooksey; Tanja McKay

Call Number

LD 251 .A566d 2021 S78

Abstract

In this dissertation, we studied several applications of artificial intelligence applications to healthcare. In the first chapter, we examined a machine learning algorithm for classifying patients presenting to the emergency department with acute respiratory distress syndrome (ARDS). Patients presenting with this life-threatening condition require a quick and accurate assessment of whether the condition is infectious or cardiac in etiology as the treatments for these etiologies of ARDS differ significantly. We used a transfer learning approach to develop our model. The model used a combination of clinical data and a chest x-ray as its input and achieved an accuracy 0.675 on the infection label and 0.745 on the cardiac label. In the second chapter, we developed models for rapid diagnosis of COVID19 from chest x-ray images, extending our suite of diagnostic algorithms for emergency room patients. These models also used a transfer learning approach and were able to achieve high sensitivity, precision, and f1-scores for detection of COVID19. Finally, we examined an algorithm for iv segmenting kidney tumors from CT scans. This algorithm was developed during the KITS19 Grand Challenge Competition [53]. Our model performed well, achieving a Sørensen–Dice score of 0.949 [61, 62, 63]. At the time of judging, we placed 50th out of all teams competing globally and 5th out of teams from the United States.

Rights Management

This work is licensed under a Creative Commons Attribution 4.0 International License.

Recommended Citation

Stubblefield, Jonathan William, "Artificial Intelligence Algorithms for Medical Imaging and Healthcare" (2021). Student Theses and Dissertations. 282.
https://arch.astate.edu/all-etd/282