AbstractAutomating the cell detection and enumeration through a picture is a way to save time
for biologists during their experiments. However, if the cells’ background is noisy, e.g. the surface is porous, the task can become much more complex. This thesis presents an algorithm combining filters and the morphological operations of erosion and dilation to isolate cells with a visible nucleus and count them. This is done in two stages: first, separating the cells from their background, then isolating nucleoli, distinctive part of the cell’s nucleus. Mathematical morphology allows to ignore intensity variation in images and help to remove noise elements. Identified cells are notified on the original picture and their number is given to the user. This processing is done by means of Python scripts. In order to improve the usability for people unfamiliar with programming, several interface solutions were analysed. Based on prototypes, their advantages and disadvantages were highlighted.
|Date of Award||28 Aug 2018|
|Supervisor||Jean-Marie JACQUET (Supervisor)|