Javid, T: Modeling Spherical Image Objects Using Wavelets

This research project is about development of a shape modeling system for sphere like objects found in digital images. The shape models are approximations to corresponding discrete image parts or regions of interests in a digital image. A sphere is a natural settings and particularly suitable for distant image objects. However, there are many scenarios in which a sphere approximation is being used successfully for example shape of earth, human eyeball, heart, animations, and computer graphics, to name a few. In medical image processing and analysis, the left ventricle of human heart is practically modeled usually with an ellipsoid model. In this project, a 3D regular hexagonal model is developed for a better approximation of spherical shape than the ellipsoid model. The better results are achieved for left ventricle ejection fraction computation. Most of the work is completed manually as a proof of concept for the possibility of such a system for practical use. However, a real implementation of this work is underway by the author and colleagues. The queries and comments are welcome and can be directed to the author.