Segmentation of Whole Cells and Cell Nuclei From 3-D Optical Microscope Images Using Dynamic Programming

摘要: Communications between cells in large part drive tissue development and function, as well disease-related processes such tumorigenesis. Understanding the mechanistic bases of these necessitates quantifying specific molecules adjacent or cell nuclei intact tissue. However, a major restriction on analyses is lack an efficient method that correctly segments each object (cell nucleus) from 3-D images specimen. We report highly reliable accurate semi-automatic algorithmic for segmenting fluorescence-labeled images. Segmentation begins with semi-automatic, 2-D delineation user-selected plane, using dynamic programming (DP) to locate border accumulated intensity per unit length greater any other possible around same object. Then two surfaces planes above below selected plane are found algorithm combines DP combinatorial searching. Following segmentation, perceived errors can be interactively corrected. accuracy not significantly affected by intermittent labeling surfaces, diffuse spurious signals away surfaces. The unique strength segmentation was demonstrated variety biological samples where all cells, including irregularly shaped were accurately segmented based visual inspection.