Level Set Segmentation of Cellular Images Based on Topological Dependence.
ABSTRACT Segmentation of cellular images presents a challenging task for computer vision, especially when the cells of irregular shapes
clump together. Level set methods can segment cells with irregular shapes when signal-to-noise ratio is low, however they
could not effectively segment cells that are clumping together. We perform topological analysis on the zero level sets to
enable effective segmentation of clumped cells. Geometrical shapes and intensities are important information for segmentation
of cells. We assimilated them in our approach and hence we are able to gain from the advantages of level sets while circumventing
its shortcoming. Validation on a data set of 4916 neural cells shows that our method is 93.3 ±0.6% accurate.
- [show abstract] [hide abstract]
ABSTRACT: Analyzing cellular morphologies on a cell-by-cell basis is vital for drug discovery, cell biology, and many other biological studies. Interactions between cells in their culture environments cause cells to touch each other in acquired microscopy images. Because of this phenomenon, cell segmentation is a challenging task, especially when the cells are of similar brightness and of highly variable shapes. The concept of topological dependence and the maximum common boundary (MCB) algorithm are presented in our previous work (Yu et al., Cytometry Part A 2009;75A:289-297). However, the MCB algorithm suffers a few shortcomings, such as low computational efficiency and difficulties in generalizing to higher dimensions. To overcome these limitations, we present the evolving generalized Voronoi diagram (EGVD) algorithm. Utilizing image intensity and geometric information, EGVD preserves topological dependence easily in both 2D and 3D images, such that touching cells can be segmented satisfactorily. A systematic comparison with other methods demonstrates that EGVD is accurate and much more efficient.Cytometry Part A 02/2010; 77(4):379-86. · 3.71 Impact Factor
Conference Proceeding: Segmentation of Neural Stem/Progenitor Cells Nuclei within 3-D Neurospheres.[show abstract] [hide abstract]
ABSTRACT: Neural stem cells derived from both embryonic and adult brain can be cultured as neurospheres; a free floating 3-D aggregate of cells. Neurospheres represent a heterogenous mix of cells including neural stem and progenitor cells. In order to investigate the self-renewal, growth and differentiation of cells within neurospheres, it is crucial that individual nuclei are accurately identified using image segmentation. Hence effective segmentation algorithm is indispensible for microscopy based neural stem cell studies. In this paper, we present a seed finding approach in scale space to identify the center of nuclei in 3-D. Then we present a novel segmentation approach, called “Evolving Generalized Voronoi Diagram”, which uses the identified centers to segment nuclei in neurospheres. Comparison of our computational results to mannually annotated ground truth demonstrates that the proposed approach is an efficient and accurate segmentation approach for 3-D neurospheres.Advances in Visual Computing, 5th International Symposium, ISVC 2009, Las Vegas, NV, USA, November 30 - December 2, 2009, Proceedings, Part I; 01/2009
- [show abstract] [hide abstract]
ABSTRACT: Natural images are depicted in a computer as pixels on a square grid and neighboring pixels are generally highly correlated. This representation can be mapped naturally to a statistical physics framework on a square lattice. In this paper, we developed an effective use of statistical mechanics to solve the image segmentation problem, which is an outstanding problem in image processing. Our Monte Carlo method using several advanced techniques, including block-spin transformation, Eden clustering and simulated annealing, seeks the solution of the celebrated Mumford–Shah image segmentation model. In particular, the advantage of our method is prominent for the case of multiphase segmentation. Our results verify that statistical physics can be a very efficient approach for image processing.New Journal of Physics 01/2011; 13(2):023004. · 4.06 Impact Factor