Deep Learning to Classify Lung Nodule Malignancy
Brown Bag Lecture by Dr. Feng Yang | 10/24/2017 11:00AM – 12PM | 7th Floor Conference Room, Bldg 38A
Lung cancer is an aggressive disease carrying a dismal prognosis with a 5-year survival rate at 18%. Despite the development of multi-modality treatments over the past decade, lung cancer accounts for approximately 27% of all cancer deaths. We investigated the likelihood of nodule malignancy classiﬁcation using CT images and end-to-end machine-learning architectures. We proposed a multi-scale CNN model and a multi-crop CNN to automatically capture nodule heterogeneity by extracting discriminative features. Experimental results demonstrate the effectiveness of our methods in classifying malignant and benign nodules without nodule segmentation.
Dr. Yang is a Visiting Scientist at CEB. She received her PhD from National Institute of Applied Science (INSA Lyon) in France in 2011, and her B.S. and M.S. degrees from Northwestern Polytechnical University in China. From 2011 to 2012, she worked at INSA Lyon as a postdoc researcher. She is currently an associate professor in Computer Science at Beijing Jiaotong University. Her research focuses on: 1) Machine learning based disease classification/prediction; 2) cardiac image processing and analysis, including cardiac image registration, diffusion tensor interpolation, atlas construction and fiber architecture validation. She has published more than 30 papers in international journals including Medical Image Analysis, IEEE TMI, IEEE TBME, Pattern Recognition, etc., and at international conferences including IPMI, MICCAI, ISMRM. She is principal investigator for 10 grants including Chinese-French cooperative project, grants from NSFC, Ministry of Chinese Education and others.