Tessa  Gustafson

Tessa Gustafson

Scholar Title

MIT EECS | Lincoln Laboratory Undergraduate Research and Innovation Scholar

Research Title

Building CNNs to Classify Dysplastic Nevi vs. Melanoma

Cohort

2020–2021

Department

EECS

Research Areas
  • Artificial Intelligence & Machine Learning
Supervisor

Amar Gupta

Abstract

A dysplastic nevus is a type of mole that 1 in 10 Americans have. They are often irregular in shape, texture, and color. Despite their irregular appearance, dysplastic nevi are non-cancerous. Melanoma looks quite similar to dysplastic nevi, as these moles are also irregular in shape, texture, and color. Melanoma, on the other hand, is in fact cancerous. Because dysplastic nevi and melanoma often present similarly, doctors are often unsure as to which irregular moles are cancerous or non-cancerous. Because of this, doctors will take skin biopsies and then examine the results to determine whether a mole was just dysplastic nevi or melanoma. I wondered whether a model could be built to take images from skin biopsies to classify these irregular moles as cancerous or non-cancerous.

Quote

I am participating in the SuperUROP program to gain more experience in computational biological research. I became interested in this research scope due to Professor Kellis and two of his computational biology courses I took. The research I conducted in these course was focused specifically in computational approaches to solving cancer related questions. This is what inspired my research proposal for this SuperUROP.

Back to Scholars