Grid Generation and image registration
主 题: Grid Generation and image registration
报告人: Professor Guojun Liao (University of Texas at Arlington)
时 间: 2015-08-08 15:30-16:30
地 点: 理科一号楼1303
Grid (mesh) generation and adaptation is a key research area for next generation of CFD and other real world engineering simulation capacities. Image registration is the process of establishing a pixel to pixel correspondence between two “similar” images. It is a key research area for image based decease diagnosis and treatment. It has been discovered that these two seemingly disjoint areas in fact are closely linked to each other. In the first part of the talk, l describe a set of new techniques for generation of grids with prescribed Jacobian determinant and the curl vector, and techniques for adaptation of grids according to physical features of numerical solutions or geometry changes. In particular, we will show our results on generation of triangular grids (meshes) with prescribed boundary nodes and curved triangular grids. In the second part of the talk, I discuss interesting interactions between grid generations and image registrations. In particular, I will describe an optimal control approach to non-linear image registration. The approach, in return, has inspired our new variational method for generation of grids with both prescribed (discrete) Jacobian determinant and curl vector.