M.S. Project Defense: Sara Abasi
Lung Cancers: Prediction with Convolutional Neural Networks
Advisor: Dr. Yanqing Zhang
Lung cancer strikes 225,000 people every year in the United States and accounts for billions of dollars in health-care costs. Primary detection is critical to give patients the best chance at recovery and survival and earlier access to life-saving interventions. Accurate detection also lowers the false positive rate that current detection technology suffers from. This high false positive rate leads to unnecessary patient anxiety, additional follow-up imaging, and interventional treatments.
Convolutional Neural Networks (CNNs) are a deep learning architecture employed to identify the hierarchy or conceptual structure of an image and give the most accurate results in predicting whether scans are cancerous. CNNs are trainable multistage architectures, with each stage consisting of multiple layers. This project investigates the best architecture for CNNs that results in the highest prediction accuracy rate and a lower loss rate.
The data set consisted of thousands of high-resolution lung scans provided by the National Cancer Institute. Data preprocessing was done in Python and a cloud platform for designing deep learning algorithms was used.
Dr. Yanqing Zhang (chair)
Dr. Anu Bourgeois