Cytopathological Image Processing(Thesis Diploma NTUA)
In this thesis advanced algorithms about medical image segmentation were developed. These algorithms were developed in the environment of MATLAB, which is a programming environment using matrices(MATrix LABoratory). Specifically, the developed algorithms were: simple algorithms for filtering and smoothing the initial image using morphological operators and median filter, the Otsu's method for thresholding, the color-based k-means algorithm. It was given special attention to morphological segmentation and specifically the watershed transform in different versions. Finally an algorithm based on distance transform was developed and there was extraction of two basic cyto image characteristics, the area and the roundness of each cell.