AbstractsEngineering

3D DEFORMABLE CONTOUR SURFACE RECONSTRUCTION: AN OPTIMIZED ESTMATION METHOD

by NANDINI MUKHERJEE




Institution: University of Cincinnati
Department: Engineering : Electrical Engineering
Degree: MS
Year: 2004
Keywords: Image Segmentation; Region growing; Level Set; Fast March algorithm; 3D contour deformation; snakes
Record ID: 1743842
Full text PDF: http://rave.ohiolink.edu/etdc/view?acc_num=ucin1078255615


Abstract

The object of this thesis is to develop an application for three-dimensional (3D) surface reconstruction from 3D point data. This development is to extend a particular newly derived 2D deformable contour extraction method to 3D surface reconstruction. Initially the functional flowchart is designed implementing the algorithm into task modules. Each task modules are coded and subsequently all the modules are put together to form the final software. The software package is developed using Microsoft Visual C++. Experiments were performed on test data, some mathematically generated and others obtained from MRI scan. The first set of mathematically generated 3D data consists of a 3D solid sphere of uniform gray scale value surrounded by a boundary of a different uniform grayscale value. The second set of mathematical data is a hollow cylinder of inner radius 5 and outer radius 15.The third set of data consists of 20 sequential slices of fat data from stomach obtained from MRI scans. In all cases the application tested favorably.