Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
BP: Fachverband Biologische Physik
BP 15: Postersession III
BP 15.19: Poster
Dienstag, 13. März 2018, 14:00–16:00, Poster B
Segmentation of dark field images from scanning X-ray micro-diffraction — •Chiara Cassini1, Andrew Wittmeier1, Manfred Burghammer2, and Sarah Köster1 — 1Institute for X-Ray Physics, Georg-August-Universität Göttingen, Germany — 2ESRF, Grenoble, France
Imaging nanostructures within cells presents several challenges: a high resolution method, capable of retrieving structural information at subcellular length scales, without the need for slicing the cells, is preferable. Optical imaging techniques and electron microscopy meet some, but not all of this requirements. Due to their small wavelength and high penetration depth, X-rays can access the nanometer range in intact cells. In particular, we focus on scanning micro-diffraction. Our samples are freeze-dried cells grown on SiN windows; each window contains several hundreds of cells. In the past, each cell scan took minutes to hours. However, we have recently employed a special fast scanning mode that allowed us to image an entire window within a single scan, in approximately 7 hours only. This approach ensures the collection of a statistically significant pool of data in a realistic timespan. However, the data analysis becomes more challenging: the selection of the different regions of interest to be analyzed is usually performed by hand on the dark field image of a single scan, but this is not feasible on images containing hundreds of cells. Thus, an automated alternative is required. A semi-automated segmentation strategy, based on Bradley's and Otsu's thresholding, is presented for the selection of background, cytoplasm and nuclei regions.