Date of Award

Summer 8-22-2025

Level of Access Assigned by Author

Open-Access Thesis

Degree Name

Master of Science in Biomedical Engineering

Department

Biomedical Engineering

First Committee Advisor

Caitlin Howell

Second Committee Member

Mehdi Tajvidi

Third Committee Member

William DeSisto

Abstract

Mycelial biocomposite materials show great promise as sustainable alternatives for a variety of applications, however these materials can take weeks to grow and are difficult to quantify. This can pose an issue in meeting production demands and ensuring consistent properties. The purpose of this work is to explore culturing and quantification methods that allow for the growth of mycelial materials to be enhanced and analyzed. To do this, a variety of culturing methods were evaluated including the addition of bacterial and soil filtrates. While bacteria filtrates inhibited the growth of mycelial materials, soil filtrates showed promise in increasing the growth rate. Fertilizer was tested to increase the growth rate in a more controlled manner, however mycelial biocomposite samples were difficult to compare due to irregular growth patterns which made visual assessment difficult. To better quantify the growth of samples in a nondestructive manner, we compared several methods of analyzing images of the mycelial material surface. This included binary image analysis, manual segmentation, semi-automated segmentation, and a machine learning algorithm. The results showed that binary image analysis methods were able to identify growth and provide quantitative growth data, however they did not provide information on growth density across the samples. While manual segmentation methods were able to identify density differences while still providing quantitative data, they were time-consuming and lacked fine detail. Semi-automated segmentation allowed for this detail to be recovered, however the generated masks showed variation between individuals. A machine learning algorithm was then trained based on the semi-automated masks and was able to generate masks efficiently and with high accuracy. Our results demonstrate that the need for a growth quantification method for mycelial materials can be addressed using image analysis, segmentation methods, and machine learning algorithms to provide growth data.

Download

Share