Machine learning provides methods, techniques, and tools that can help solving diagnostic and prognostic problems in a variety of medical domains. In this talk, I will present our recent work of developing machine-learning algorithms in medical imaging computing applications. In the first part, I will introduce how to segment different brain tissues on first-year infant MR images by using random forest technique with multimodality features. In the second part, I will demonstrate how to enhance image resolution of routine 3T MR images towards 7T image resolution and contrast by using paired dictionaries and sparse models. In the last part, I will talk about computer-aided prediction of autistic risk on 6-month infants by using multi-kernel SVM with features from brain white matter fiber tracts.
Dr. Shi Feng is an Assistant Professor in the Dapartment of Radiology at the University of North Carolina at Chapel Hill, USA. He received his Bachelor’s Degree in electronics engineering from Peking University in 2002, and obtained his PhD in Computer Science from Institute of Automation, Chinese Academy of Sciences in 2008. He subsequently obtained postdoctoral training in Medical Image Computing at the UNC, where he has been a Research Assistant Professor of Radiology since 2011. He is also an affiliated faculty with the UNC Biomedical Research Imaging Center (BRIC). His research interests include pattern recognition and machine learning, multimodal medical image analysis, neuroanatomy and cognitive neuroscience, and statistical applications in neurodevelopment, neurological and psychiatric diseases.