Abstract by Wilson Redd

Personal Infomation

Presenter's Name

Wilson Redd


Spring Cullen

Degree Level


Abstract Infomation


Computer Science

Faculty Advisor

David Wingate


But how do you detect a cancerous cell?


Cervical cancer is a dangerous and prevalent disease throughout the world. In India, the situation for women is particularly severe. Cervical cancer accounts for 26% of cancer for women in India, with only 20% having been screened at least once in their lifetime; comparatively, 70% of American women see a gynecologist every few years.


In order to improve medical treatment of women in regions like India, our research focuses on increasing rates of early detection of cancerous cells from PAP-smear samples. Since roughly 90% of slides collected in India are non-cancerous, we are applying deep learning to visually examine and triage potentially cancerous slides, thereby enabling pathologists to spend more time examining abnormal slides.


In this presentation, we put forth a Cancerous MNIST dataset we have created to replicate slide conditions and outline the pipeline and machine learning algorithms we will be developing over the coming year.