Segmentation and tracking of neural progenitor cells in microscopic image sequences |
|
| Abstract | The goal of the thesis is to present a pipeline for automatic segmenting and tracking of cells in phase-contrast microscopy image sequences. The result from such a pipeline is a number of graphs showing the lineage of the cells from the initial image frame and their development as it progresses over time. The collected data contain information on position, size and speed of each cell and also frequency of mitosis for each cell. This is useful for research in stem cells and the development hereof. This is an essential tool for registering and comparing the effects of different treatments of the cell cultures.
The segmentation and tracking is based on level-set theory and implemented as a modified version of the active contour formulation as proposed by Chan and Vese [CV01]. In the active contour model a level set of an implicit function is used to de fine the contour. The cells can not merge and therefore the model also needs to be constrained to avoid the individual segments merge.A cost-function is constructed which is minimized through a series of iterations between each frame. The active contour model adapt the contour to the shape and movement of the cell between each frame.
The model is applied to the dataset for validation purposes and for comparison of performance . The dataset used is fully manually annotated and therefore suited for validation and benchmarking. |
| Type | Master's thesis [Academic thesis] |
| Year | 2013 |
| Publisher | Technical University of Denmark, DTU Compute, E-mail: compute@compute.dtu.dk |
| Address | Matematiktorvet, Building 303-B, DK-2800 Kgs. Lyngby, Denmark |
| Series | M.Sc.-2013-65 |
| Note | DTU supervisors: Rasmus Larsen, rlar@dtu.dk, and Knut Conradsen, knco@dtu.dk, DTU Compute |
| Electronic version(s) | [pdf] |
| Publication link | http://www.compute.dtu.dk/English.aspx |
| BibTeX data | [bibtex] |
| IMM Group(s) | Image Analysis & Computer Graphics |