Analysis of Microscopic Images

Analysis of Microscopic Images: A Morphological Approach

van Belkom, Myrte (TU Delft Electrical Engineering, Mathematics and Computer Science)

Budko, Neil (mentor)
den Ouden-van der Horst, Dennis (graduation committee)
van Elderen, Emiel (graduation committee)

Delft University of Technology

2018-06-27

In this thesis we apply the numerical method of Morphological Geometric Active Contours as proposed by Alvarez, Baumela and Marquez-Neila to microscopic images of plant cells. The goal is to find all plant cells in the images, and then to find their cell walls. This is done in order to calculate certain properties of these cells automatically, such as area, perimeter and ellipticity. As the name suggests, mathematical morphology plays a big part in morphological GAC. This thesis describes the theory behind morphological GAC, mathematical morphology and the implementation in Python of the actual algorithm. We made two adaptations compared to the original model. These are a change in the conditions of the balloon force, and an extra step at the end of each iteration. The change in the balloon force was made to keep edges stronger. The extra step was added to improve smoothness of the contours, since the smoothing parameter had no effect.

Morphology
Geometric Active Contours
Numerical
Python

http://resolver.tudelft.nl/uuid:6c58e3ea-d9e5-4a95-9552-95ca8576b035

Student theses

bachelor thesis

© 2018 Myrte van Belkom