Machine Learning Seminar
Speaker: Ge Wang, Director, Biomedical Imaging Center, Rensselaer Polytechnic Institute
Title: Integrated Imaging in the Machine Learning Framework
Date: Mon, 15 Jan 2018, 2:15 pm – 3:15 pm
Location: 1400 BPS Bldg.
Computer vision and image analysis are both great examples of successes with machine learning especially deep learning. While computer vision and image analysis primarily deal with existing images, tomographic reconstruction produces images of internal structures from externally measured indirect data. Recently, deep learning techniques are being actively developed for tomographic reconstruction by multiple groups worldwide, with encouraging results at RPI and other institutions. We believe that deep reconstruction is a next major target of deep learning, has a revolutionary potential to improve tomography and solutions to other inverse problems, and promises major impacts on imaging. Along this direction, we have been working on data-driven design of tomographic algorithms, for optimized workflow with multi-stages and/or multi-modes, and toward superior performance in important biomedical applications. In addition to a general perspective (http://ieeexplore.ieee.org/document/7733110), several “deep imaging” projects of ours will be reported, involving CT, MRI, radiomics, and multimodality imaging. Some methodological explorations will be also discussed, including 2nd-order neural networks and so on.
Ge Wang is the Clark & Crossan Endowed Chair Professor and the Director of the Biomedical Imaging Center, Rensselaer Polytechnic Institute, Troy, NY, USA. He authored the pioneering paper on the first spiral/helical cone-beam/multi-slice CT algorithm in 1991. Currently, there are over 100 million medical CT scans yearly with a majority in the spiral/helical cone-beam/multi-slice mode. He and his collaborators published the first paper on bioluminescence tomography. His group published the first papers on interior tomography and omni-tomography (“all-in-one”) to acquire diverse datasets simultaneously (“all-at-once”) with CT-MRI as an example. His results were featured in Nature, Science, and PNAS, and recognized with academic awards. He wrote over 450 peer-reviewed journal papers, receiving a high number of citations. His team has been in collaboration with world-class groups and continuously well-funded. His interest includes X-ray CT, optical molecular tomography, multimodality fusion, and machine learning for medical imaging (supported by GE Global Research Center). He is the Lead Guest Editor of the five IEEE Transactions on Medical Imaging Special Issues, the founding Editor-in-Chief of the International Journal of Biomedical Imaging, and an Associate Editor of the IEEE Transactions on Medical Imaging, Medical Physics, and other journals. He is a fellow of the IEEE, SPIE, OSA, AIMBE, AAPM, and AAAS.