Thumbnails:
List:
Year:
Category:
Session:
Poster:
Getting poster data...
Britta Weber, Marit Mller, Steffen Prohaska, David Gnther, Hans-Christian Hege (Zuse Institute Berlin, Berlin, Germany)
The cytoskeleton, in particular the microtubule network, is currently one of the most important research topics in cell biology. The study of this network involves image processing to extract information about length, number, distribution, and type of the microtubules. One popular method for image acquisition is electron tomography which can reveal not only the three-dimensional structure of the microtubule network at a nanometer scale but also allows classification of the microtubules ends. In order to draw reliable conclusions about the structure of microtubule networks, a large number of microtubules needs to be precisely identified. However, while manual segmentation of all microtubule centerlines is very cumbersome and often not feasible, automatic extraction is challenging and remains an open problem. We present an editor specifically designed for the task of segmentation and classification of microtubules. Experimental results evaluating a semi-automatic workflow suggest that a well-designed tracking algorithm and specific enhancements of the user interface along with a useful visualization can speed up manual segmentation and classification of microtubules by a factor of 2.