Integrative Sparse Modeling and Classiﬁcation of Biomedical Imaging Patterns
Brown Bag Lecture by Dr. Keni Zheng | 11/6/2018 11:00AM – 12PM | 7th Floor Conference Room, Bldg 38A
The analysis and characterization of imaging patterns are signiﬁcant for biomedicine and numerous other domains. Here we develop mathematical methods and algorithms for computer aided-diagnosis. The central hypothesis is that we can predict the occurrence of diseases using supervised learning techniques applied to medical image datasets that include healthy and diseased subjects for training. We develop methods to calculate sparse representations to classify imaging patterns and compare these with traditional texture-based classiﬁcation. We also explore dictionary learning techniques for possible improvement of classiﬁcation accuracy. The two application domains are osteoporosis diagnosis in radiographs of the calcaneus bone, and breast lesion characterization in mammograms, both important for public health. Our proposed classiﬁcation system shows signiﬁcant improvement for these two data sets.
Dr. Zheng earned her Ph.D. in Applied Mathematics at Delaware State University (DSU) recently. She is focused on analysis of medical images by machine learning and pattern recognition. She has developed mathematical methods and algorithms for learning, and created models that recognize diseases at an earlier stage. The applications she is working on are breast cancer and osteoporosis.