|Céline Di Venuto||Semester project|
|Section Microtechnique, EPFL||February 2007|
The aim of this project is to evaluate the growth of blood vessels in the vascular model chloro-allantoic of the chicken embryo. The goal of this project is to develop an image-processing algorithm, which allows a quantitative measure of the temporal evolution of the vascular density. The analysis is done on time-lapse microscopic images. Two distinct measures were developed. The first one allows the user to specify a starting direction on a vessel. Then the program tries to automatically trace the vascular tree. The second measure has two cases; the first one consists in computing the density of vessels in a region of interest specified by the user. In the second case, it computes the density and orientation in the whole image, which is segmented into squares of a certain size. It also computes a measure of validity for the orientation based on the eccentricity of the best ellipsis in the image. The minimal eccentricity as well as the size of the squares can be chosen by the user. The two measures are based on some pre-processing steps which remove the non-uniform background in images and on some processing steps which helps to find the vessels in the image thanks to steerable filters.