Monday, March 12
Probabilistic Machine Learning for Lesion and Tumour Detection, Segmentation and Disease Prediction in Patient Brain Images.”
Tal Arbel’s research focuses on developing probabilistic machine learning techniques in computer vision for medical image analysis, with a wide range of applications in neurology and neurosurgery. She will describe recent work developing probabilistic graphical models for brain tumour/lesion detection and segmentation, which were successfully applied to the MICCAI BRaTs brain tumour segmentation challenge datasets and to large-scale, multi-scanner, multi-center clinical trial datasets of patients with Multiple Sclerosis. Additional graphical models were developed for accurate detection and segmentation of active lesions in contrast-enhanced images, and for new lesions in longitudinal patient MRI acquired over several timepoints, both of which are important markers of new disease activity and for assessing treatment effects in clinical trials. Tools developed in her lab have been integrated into the software analysis pipeline of an industrial partner for usage in clinical trial drug development., where the methods have assisted in the analysis of almost all the new MS treatments currently being used worldwide. She will describe recent work for the prediction of future new lesion activity based on baseline MRI, and for automatically identifying potential responders to treatment, leading to the possibility of personalized medicine.
Tal Arbel is a Professor in the Department of Electrical and Computer Engineering, and the Director of the Probabilistic Vision Group and Medical Imaging Lab in the Centre for Intelligent Machines, McGill University, Montréal, Canada. Her expertise lies in the development of probabilistic and machine learning methods in computer vision in the context of medical image analysis, particularly in neurology and neurosurgery. Prof. Arbel has particularly extensive expertise in developing probabilistic models for brain tumour/lesion detection and segmentation. She has developed models for computational neuroanatomy, with the objective of generating automatic discoveries of healthy brain morphometry. Multi-modal image registration tools developed in her lab have been integrated into the operating theatres of the Montreal Neurological Hospital, where they are being used to assist in image-guided neurosurgery for tumour resections. Her lab is currently developing machine learning methods for the automatic identification of biomarkers of neurodegenerative disease progression. Prof. Arbel has authored close to 100 peer-reviewed papers and has co-organized a number of major international conferences in both fields, including serving as co-organizer and satellite events chair for MICCAI 2017, area chair/program committee member for CVPR and MICCAI, and General Chair for a major joint national conference (AI/GI/CRV/IS). She is currently an Associate Editor for IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) and the Journal of Computer Vision and Image Understanding (CVIU).
Tuesday, March 13
Materials in the wild: recognition, editing, and stylization
A rich range of materials contribute to the visual appearance and aesthetics of the environments we live in. But materials are often not treated as first-class citizens, even though consumers have strong preferences on visual appearance and style that drive their purchasing decisions. In this talk I will describe our work on material recognition in the wild, fine-grained product recognition, material editing, and image stylization. This research has broad applications in e-commerce and retail, in virtual and augmented reality, and in industrial and interior design.
Kavita Bala is a Professor in the Computer Science Department at Cornell University and Chief Scientist at GrokStyle. She is the Editor-in-Chief of Transactions on Graphics (TOG). Her research projects have been commercialized into Autodesk’s production cloud renderer and GrokStyle’s visual search engine; and her work on 3D Mandalas was featured at the Rubin Museum of Art, New York.
Wednesday, March 14
Challenge problems within computer vision play a critical role in advancing research. IARPA has led or is leading challenge problems looking into face recognition, fingerprint capture, facial feature fusion, facial disguise, and improving UAV image quality. This talk will provide an overview of the importance of prize challenges and results for some of the different challenges I have run.
Dr. Chris Boehnen is a Senior Program Manager at the Intelligence Advanced Research Projects Activity (IARPA) focused on biometrics, computer vision, and machine learning. He is the PM for the Odin, Janus, N2N Challenge, and BEST programs. He is also joint faculty at the University of Tennessee. Dr. Boehnen was formerly the founder and team lead for the Secure Computer Vision Team at Oak Ridge National Laboratory (ORNL). In his six years at ORNL he served as Principal Investigator on $11 million in funding spread over 24 different grants which he conceived, proposed, and successfully executed. He received the ORNL Early Career Award for Engineering and 3 of his papers have received best paper awards at highly competitive conferences including best paper out of 133 submissions at BTAS 2016. Dr. Boehnen received his B.S., M.S. and Ph.D. from the University of Notre Dame Computer Science and Engineering Department. He has been a member of the biometrics research community since 2001 when he began working on the Face Recognition Grand Challenge.