Year | Month |
2023 | Jan Feb Mar Apr May Jun Jul Aug Oct |
2022 | Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec |
2021 | Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec |
2020 | Jan May Jun Jul Aug Sep Oct Nov Dec |
2019 | Jan Feb Mar Apr May Jun Jul Aug Oct Nov |
Title: | Multi-tasking Inverse Problems: More Together Than Alone |
Speaker: | Prof Carola-Bibiane Schönlieb, University of Cambridge, United Kingdom |
Time/Place: | 16:00 - 17:00 Zoom, (Meeting ID: 991 4880 3896) |
Abstract: | Inverse imaging problems in practice constitute a pipeline of tasks that starts with image reconstruction, involves registration, segmentation, and a prediction task at the end. The idea of multi-tasking inverse problems is to make use of the full information in the data in every step of this pipeline by jointly optimising for all tasks. While this is not a new idea in inverse problems, the ability of deep learning to capture complex prior information paired with its computational efficiency renders an all-in-one approach practically possible for the first time. In this talk we will discuss multi-tasking approaches to inverse problems, and their analytical and numerical challenges. This will include a variational model for joint motion estimation and reconstruction for fast tomographic imaging, joint registration and reconstruction (using a template image as a shape prior in the reconstruction) for limited angle tomography, as well as a variational model for joint image reconstruction and segmentation for MRI. These variational approaches will be put in contrast to a deep learning framework for multi-tasking inverse problems, with examples for joint image reconstruction and segmentation, and joint image reconstruction and classification from tomographic data. |
Title: | Learning Geometry |
Speaker: | Professor Ron Kimmel, Technion - Israel Institute of Technology, Israel |
Time/Place: | 16:00 - 17:00 Zoom, (Meeting ID: 914 4896 2795) |
Abstract: | Deep learning is a disruptive line of research that changes the way computational problems are being addressed and solved. Many parameters are optimized for, and tuned to train a given computational architecture to classify, segment, identify, and reconstruct objects. This methodology works great as long as there is some assumption about the size of the data, or its spatial or temporal shift invariance property that allow convolutional neural networks to operate on the given data. The question we will address is what can be done for geometric structures for which there is no linear shift invariance mechanism to rely on. We will relate to matching geometric structures, measuring geodesic distances, classifying objects, and ways to import axiomatic constructions into the learning arena, that give birth to novel semi-supervised learning procedures. Finally, I will comment on a new and interesting line of research we have started to explore dealing with computational pathology. |
We organize conferences and workshops every year. Hope we can see you in future.
Learn MoreProf. M. Cheng, Dr. Y. S. Hon, Dr. K. F. Lam, Prof. L. Ling, Dr. T. Tong and Prof. L. Zhu have been awarded research grants by Hong Kong Research Grant Council (RGC) — congratulations!
Learn MoreFollow HKBU Math