Return to search

Development of a semi-automatic method for cellular migration and division analysis

Binary image processing algorithms have been implemented in this study to create a background subtraction mask for the segmentation of cellular time lapse images. The complexity in the development of the background subtraction mask stems from the inherent difficulties in contrast resolution at the cellular boundaries. Coupling the background subtraction mask with the path reconstruction method via superposition of overlapping binary segmented objects in sequential time lapse images produces a semi-automatic method for cellular tracking. In addition to the traditional center of mass or centroid approximation, a novel quasi-center of mass (QCM) derived from the local maxima of the distance transformation (DT) has also been proposed in this study. Furthermore, image isolation and separation between spreading/motile and mitotic cells allows the extraction of both migratory and divisional cellular information. DT application to isolated mitotic cells permits the ability to identify distinct morphologic phases of cellular division. Application of standard bivariate statistics allows the characterization of cellular migration and growth. Determination of Hotelling???s confidence ellipse from cellular trajectory data elucidates the biased or unbiased migration of cellular populations. We investigated whether it was possible to describe the trajectory as a simple binomial process, where trajectory directions are classified into a sequence of (8) discrete states. A significant proportion of trajectories did not follow the binomial model. Additionally, a preliminary relationship between the image background area, approximate number of counted cells in an image frame, and imaging time is proposed from the segmentation of confluent monolayer cellular cultures.

Identiferoai:union.ndltd.org:ADTP/258527
Date January 2005
CreatorsChu, Calvin, School of Biomedical Engineering, UNSW
PublisherAwarded by:University of New South Wales. School of Biomedical Engineering
Source SetsAustraliasian Digital Theses Program
LanguageEnglish
Detected LanguageEnglish
RightsCopyright Calvin Chu, http://unsworks.unsw.edu.au/copyright

Page generated in 0.0014 seconds