Membrane boundary extraction using circular multiple paths
Membrane proteins represent over 50% of known drug targets. Accordingly, several widely used assays in the high content analysis area rely on quantitative measures of the translocation of proteins between intracellular organelles and the cell surface. In order to increase the sensitivity of these assays, one needs to measure the signal specifically along the membrane, requiring a precise segmentation of this compartment. Manual tracing of membrane boundary is very time-consuming and confronts us with issues of objectivity and reproducibility. In this paper, we present an approach based on a circular multiple paths technique on transformed images that enables us to segment the membrane compartment accurately and rapidly. We have presented three approaches for image transformation. The circular property of the multiple paths ensures that we are obtaining closed contours for the membrane boundary. The position of the multiple paths provides the edges of the membrane boundary. The effectiveness of our algorithm is illustrated using cells expressing epitope-tagged membrane proteins.
|Authors||Sun, C.; Vallotton, P.; Wang, D.; Lopez, J. A.; Ng, Y.; James, D. E.;|
|Publisher Name||PATTERN RECOGNITION|
|OpenAccess link to author's accepted manuscript version||https://publications.gimr.garvan.org.au/open-access/10470|