Tuesday, October 10th, 2017. Time: 15h.
Place: O.C. Zienkiewicz, C-1 building, Campus Nord UPC
In order to correctly model physical problems, such as natural disasters or cardiovascular pathologies, it is very important to obtain a good discretization of the domain to study. In this work we intend to reconstruct an object / domain from generated 2D projections obtained from satellite images, RX angiography or even stereographic images (mobile phones). To reconstruct this object from the 2D projections, filtering and image processing algorithms will be applied: erosion, dilation, Otsu filter, skeletonization, epipolar treatment of geometry, among others. This process will allow us to obtain different approximations of the domain: 1D model, 2D model and 3D model, depending on the computational model (FEM) that we want to apply later. This technique can be applied both to the FEM study of river flows through satellite imagery and to flow in arteries using RX angiography images.
M.C. Fernando Cervantes Sánchez from Centro Investigación de Matemáticas de México (CIMAT) - Collaboration project CIMAT-CIMNE