Biological model representation and analysis

Abstract

In this thesis, we discuss solutions of phenotype description based

on the microscopy image analysis to deal with biological problems both

in 2D and 3D

space. Our description of patterns goes beyond conventional features and helps to visualize the unseen in feature dataset. These solutions share several common processes which are based on similar principles. Furthermore, we notice that advanced features and classier strategies can help us improve the performance of the solutions. The biological problems that we have studied include the endocytosis routing using high-throughput screening in 2D and time and 3D geometrical representation from biological structures.