Yara Zayed, a student in the Data Science and Business Analytics master’s program has defended her thesis titled: “A System for Detecting Lung Cancer Using Medical Image Processing, Machine Learning and Deep Learning Algorithms.”
Through her study, the researcher presented two distinct models to facilitate the diagnosis of lung cancer using demographic data and medical images. This is done through the use of Machine Learning and Deep Learning algorithms.
In the machine learning model, the researcher used a wide range of demographic data and risk factors related to lung cancer, as the data formed the basis for training several machine learning models including XGB, RF, SM, DT, GBM, and AdaBoost. The researcher used various techniques to conduct data pre-processing represented in increasing the sample size for the data. The XGB model achieved remarkable achievement with an accuracy rate of 99.07%.
As far as the deep learning model is concerned, it took advantage of a global dataset containing CT scan images of lung cancer that were accurately classified into three categories: normal, benign, and malignant. The images were exposed to a set of careful preparation stages in terms of data augmentation and division, during which eight deep learning algorithms were trained, including CNN, MobileNet, Xception, DenseNet, VGG16, VGG19, and Efficient Net. The researcher approached the maximum performance score using the CNN model, with an exceptional accuracy rate of 99.70%, and 100% for accuracy, recall and F1 score.
The thesis was supervised by Professor Dr. Mohammed Awad. The committee of examiners included: Dr. Mahmoud Obaid and Dr. Yousef Daraghmeh.