Full text
| |
| Author(s): |
Cícero Ribeiro de Lima
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | São Paulo. |
| Institution: | Universidade de São Paulo (USP). Escola Politécnica (EP/BC) |
| Defense date: | 2006-07-14 |
| Examining board members: |
Emilio Carlos Nelli Silva;
Marcelo Britto Passos Amato;
Raul Gonzalez Lima;
Gláucio Hermogenes Paulino;
Marcos de Sales Guerra Tsuzuki
|
| Advisor: | Emilio Carlos Nelli Silva |
| Abstract | |
The Electrical Impedance Tomography (EIT) is a recent monitoring technique on biological tissues. The EIT allows us to obtain images representing a transversal plane of any section of human body (head, thorax, thigh, etc). Each image pixel is related to its corresponding value of electrical impedance (or resistivity). The images are generated from voltage values measured on electrodes positioned around the section of human body. These voltages are obtained by applying to the electrodes an alternated sequence of low intensity electrical currents in according to an excitation pattern (adjacent or diametral). The EIT is based on an inverse problem, where given the voltages measured outside of body, this technique tries to find the conductivity distribution inside of the body. In this work, the main objective is to apply Topology Optimization Method (TOM) to obtain images of body section in EIT. Topology Optimization seeks a material distribution inside of a design domain, determining which points of space should be solid and which points should be void, to minimize (or maximize) an objective function requirement, satisfying specified constraints. In this work, the MOT is an iterative method whose computational algorithm (implemented in C language) combines Finite Element Method (FEM) and an optimization algorithm called Sequential Linear Programming (SLP). The topology optimization problem applied to obtain images consists of finding the material (or conductivity) distribution in the body section that minimizes the difference between electric potential measured on electrodes and electric potential calculated by using a computational model. The main advantage of TOM applied to image reconstruction in EIT is to allow us to include several constraints in optimization problem, which reduces the solution space and avoids images without clinical meaning. In this work, the MOT uses a material model based on SIMP to makes relaxation of topology optimization problem and several regularization schemes are implemented to solve inverse problem of EIT, such as image tuning control, weighted distance interpolation based on average conductivity of domain, spatial filtering technique for gradient control, graduated changing in penalty factor of material model during the optimization process (continuity method), and continuous approximation of material distribution (CAMD). This work belongs to a thematic project whose aim is to study reconstruction image algorithms that could be used in an EIT device to monitor accurately mechanical ventilation of lung and to diagnose when any portion of lung is damaged (obstructed or collapsed) during mechanical ventilation process. To illustrate the implementation of the method, image reconstruction results obtained by using voltage numerical and experimental data of well-know 2D domains are shown. (AU) | |