טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentInberg Miriam
SubjectQuantitative Analysis of Blood Vessel Development in Medical
Images
DepartmentDepartment of Applied Mathematics
Supervisors Professor Alfred Bruckstein
Dr. Eugene Walach


Abstract

In this work a problem of medical image processing is considered. The thesis deals with the quantitative analysis of images obtained from angiogenesis experiments performed using the chick chorioallantoic membrane (CAM) angiogenesis assay. During angiogenesis new blood vessels grow from a pre-existing membrane vasculature in response to external stimuli of growth factor (bFGF).


The CAM is a growing, developing structure with vessels differing in their size, length, density and blood flow direction. The estimation of new blood vessels development includes estimation of the direction of the growth and its quantity. The geometrical shape of a vessel strongly depends on its environment characteristics, and on location of the source of influence which causes this growth. The vessel’s ability to adapt to the source of influence is a concrete physical parameter which can be measured.


The purpose of this work is to study vessel orientation using the image processing techniques.  There are no methods that enable easy quantitation and determination of blood vessels orientation. This research offers for the first time a quantitative tool for the analysis of such angiogenesis experiments. A new technique, named orientation transform, characterizes a vessel by determining the quantity of influence of growth factor (bFGF) on the vessel.  The goal of vessel orientation transform is to define the power of influence in each point, i.e. orientation value in each point. The main idea of this transform is to find the constellation of points on the skeleton image to which a greater number of vessels will be oriented.  For this purpose, the orientation transform divides vessels into the small linear segments of 5-10 pixels and finds crossing points of their prolongations. Such linearization helps to determine the highest density of crossing points which reflects the corresponding areas of attraction.


The application of orientation transform has already been used as a standard tool in the Vessel Research Laboratory of Medicine faculty at the Technion.