Major: Biomedical Engineering
Minor: Engineering Project Management
Class Year: 2026
High School: Midway High School
Advisor: Dr. Kenneth Hoyt
URS Thesis Title: Machine Learning for Estimation of Ultrasound Scatterer Size
Research Focus: Medical Imaging, Machine Learning, Ultrasound, Artificial Intelligence
Other Projects and Publications: Ensemble Deep Learning for Tissue Segmentation in Breast Cancer Histopathology
Organizations and Programs: Undergraduate Research Scholars, Undergraduate Research Ambassadors, Engineering Honors, Biomedical Engineering Society
Awards and Distinctions: President's Endowed Scholar, Dean's Honor Roll, 2024 Biomedical Engineering Society Presenting Author
Experience: Abstract Presentations, NIH REU
Years Experience: 2
Other Activities: A&M UMC Choir, Soccer Official, Hiking
Bio: Howdy! My name is Ashlyn Melichar, and I am a senior biomedical engineering student from Waco, Texas. Upon graduation, I hope to attend graduate school and earn a PhD in biomedical engineering. I plan to pursue a career in research, and I would specifically like to work in the cancer imaging and computational diagnostic field.
In my time at Texas A&M, I have worked with Dr. Kenneth Hoyt’s computational ultrasound lab to complete a thesis through the Undergraduate Research Scholars program. The goal of my project was to develop machine learning models for predicting acoustic scatterer size in ultrasound images. This parameter can be used in tissue characterization applications for diagnosing and monitoring conditions like liver disease and breast cancer. Ideally, more accurate acoustic scatterer size estimation will allow physicians to diagnose diseases like cancer sooner and monitor patient response to treatment more closely over time.
Outside of A&M, I have also had the opportunity to complete an NIH REU program through Wake Forest University’s Center for Artificial Intelligence Research. Under the mentorship of Dr. Metin Gurcan and Dr. Onur Koyun, I utilized deep learning techniques to automate tissue segmentation in breast cancer tumors. Ideally, these automated models will help pathologists to accurately distinguish tissue boundaries in breast cancer tumors and make predictions about patient outcomes.