A Unified Framework for Automated 3-D Segmentation of Surface-Stained Living Cells and a Comprehensive Segmentation Evaluation

作者：Hodneland Erlend; Bukoreshtliev Nickolay V; Eichler Tilo W; Tai Xue Cheng; Gurke Steffen; Lundervold Arvid^*; Gerdes Hans Hermann

来源：IEEE Transactions on Medical Imaging, 2009, 28(5): 720-738.

DOI：10.1109/TMI.2008.2011522

• Automated whole cell segmentation
• level set
• mathematical morphology
• region differencing
• segmentation evaluation
• watershed

This work presents a unified framework for whole cell segmentation of surface stained living cells from 3-D data sets of fluorescent images. Every step of the process is described, image acquisition, prefiltering, ridge enhancement, cell segmentation, and a segmentation evaluation. The segmentation results from two different automated approaches for segmentation are compared to manual segmentation of the same data using a rigorous evaluation scheme. This revealed that combination of the respective cell types with the most suitable microscopy method resulted in high success rates up to 97 %. The described approach permits to automatically perform a statistical analysis of various parameters from living cells.

• 出版日期2009-5