Date of Graduation

5-2023

Document Type

Thesis

Degree Name

Bachelor of Science in Biomedical Engineering

Degree Level

Undergraduate

Department

Biomedical Engineering

Advisor/Mentor

Quinn, Kyle

Abstract

Chronic wounds affect nearly 1 out of every 50 people in the United States, decreasing quality of life and putting people with potential comorbidities at high risk for obtaining an infection. Wound healing progress in clinical settings is measured by tracking wound size and there are currently no non-invasive, quantitative measurement techniques. Previous studies have proposed nicotinamide adenine dinucleotide (NADH) autofluorescence lifetime (FLIM) imaging of the wound edge as a method to quantify the wound healing process by connecting the role of NADH in cellular metabolism to wound healing stage. However, evaluation of FLIM images is heavily subjective, as a user must trace the regions of interest within an intensity image, which is time consuming, taking minutes for every picture in a set of hundreds. Because of the time constraining bottleneck, there is a need to develop an automated way to generate image masks that allow computation of FLIM parameters within tissue regions containing live cells. A convolutional neural network was trained to segment the epithelium within FLIM images by inputting ground truth user-drawn masks, and a dataset of 512x512 pixel FLIM intensity images (n=200). The accuracy of the network was confirmed by analyzing the intensity images from an independent testing set and calculating the true positive rate and true negative rate achieved by the network as well as the accuracy and F1 score. For further confirmation of a viable network, the A1% and mean lifetime (Tm) values were calculated and compared using both the ground truth and network masks, and showed no statistically significant differences between the outputs of the ground truth and network masks.

Keywords

machine learning; artificial intelligence; FLIM; metabolism

Available for download on Sunday, April 26, 2026

Share

COinS