Ali Raufi

I am currently a sophomore here at DePaul majoring in Computer Science with a minor in math. I enjoy taking what I learn in the classroom and in my free time to help be a part of creating meaningful work. I am interested in learning more about computer vision and machine learning. In my free time, I like going to the gym, drawing, and listening to music.

Comb BS in Data Science

Ilyas Ustun

ARAUFI@depaul.edu

Understanding the Relationship between Semantic Characteristics for Computer-Aided Diagnosis on Lung Nodules

Abstract: Lung cancer is a leading cause of cancer death in both men and women, making up almost 25% of all cancer deaths. While significant progress has been made to advance binary computer aided diagnosis (CAD) systems (highly malignant versus highly benign), there is still a need to predict the degree of malignancy as the hard to classify cases are the ones that are typically uncertain (somehow benign, indeterminant, somehow malignant). We hypothesize that by integrating domain knowledge, such as semantic characteristics used by medical experts to describe the visual appearance of lung nodules, the performance of the CAD systems can be furthered improved. Using the NIH/NCI Lung Image Database Consortium (LIDC) image collection, we show that there is a strong correlation between semantic characteristics and malignancy. In our preliminary results, we found that there is a high correlation between spiculation and malignancy using Kendall’s Tau-B, Yule’s Q, Goodman & Kruskal’s Gamma, and Spearman’s Rho correlation methods. Establishing the relationship between the semantic characteristics is critical since this additional information can be used to further improve the overall CAD performance.