Goal: Image processing, segmentation and statistics on in-vivo retinal pigment epithelium cells’ images.
We developed at LAPD a new technology for imaging the back of the eye and observe cellular details associated with diseases. Following the first-in-human study, a larger clinical study over 100 patients will start mid 2020 at Jules Gonin eye hospital in Lausanne on healthy volunteers and patients. Within the study, a database of retinal pigment epithelium cells images will be created.
A first image-processing algorithm was developed to segment the cells on the images enabling the computation of quantitative parameters such as cell size, density, number of neighbors, etc. However, the algorithm is not very robust and was tested only on few images.
The goal of this master thesis is to further develop the segmentation algorithm and use it on the new image database to extract cells morphological parameters.
If time is sufficient, a further goal is then to perform statistics on these parameters to highlight the differences between the groups of subject included in the clinical trial.
The work consists in the following steps:
– State of the art investigation.
– Development, optimization, test and validation of the existing MATLAB algorithm software.
– Processing of the clinical trial database.
– Statistics based on the results of the segmentation.
Project type: Master project.
– Image processing, MATLAB, Segmentation, Statistics.