Event(s) on July 2011

• 11/7/2011

  題目:	CMIV Colloquium: Semi-Supervised Subspace Learning for Mumford-Shah Model Based Texture Segmentation
  講員:	Dr. Andy Yip, Department of Mathematics, National University of Singapore, Singapore
  時間/地點:	14:30  -  15:30 DLB514, David C. Lam Building, Shaw Campus, Hong Kong Baptist University
  摘要:	We propose a novel image segmentation model which incorporates subspace clustering techniques into a Mumford-Shah model to solve texture segmentation problems. While the natural unsupervised approach to learn a feature subspace can easily be trapped in a local solution, we propose a novel semi-supervised optimization algorithm that makes use of information derived from both the intermediate segmentation results and the regions-of-interest (ROI) selected by the user to determine the optimal subspaces of the target regions. These subspaces are embedded into a Mumford-Shah objective function so that each segment of the optimal partition is homogeneous in its own subspace. The method outperforms standard Mumford-Shah models since it can separate textures which are less separated in the full feature space. Experimental results are presented to confirm the usefulness of subspace clustering in texture segmentation.

  
  

• 12/7/2011

  題目:	Efficient Algorithms for Total Variation Based Image Reconstruction and Applications in Partially Parallel MR Imaging
  講員:	Dr. Xiaojing Ye, Department of Mathematics, University of Florida, USA
  時間/地點:	11:00  -  12:00 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University
  摘要:	In this talk I will discuss several fast numerical algorithms for image reconstruction with total variation (TV) regularization. TV based image reconstruction has proved to be very promising in magnetic resonance (MR) imaging. However the solution of TV based reconstruction encounters two main difficulties on the computational aspect of the emerging MR imaging application known as partially parallel imaging (PPI): the inversion matrix is large and severely ill-conditioned, and the objective is nonsmooth. We will talk about three algorithms that tackle the problem using variable splitting, Lagrangian methods and optimal step size selection technique. The numerical results show that these algorithms outperform the state-of-the-arts in terms of effectiveness and robustness.