Unsupervised Segmentation of Overlapped Nuclei Using Bayesian Classification

• Published in: IEEE transactions on biomedical engineering Vol. 57; no. 12; pp. 2825 - 2832
• Main Authors: Jung, Chanho, Kim, Changick, Chae, Seoung Wan, Oh, Sukjoong
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01-12-2010; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Automatic cell segmentation; Bayes Theorem; Bayesian methods; Biological and medical sciences; Biomedical imaging; Breast - cytology; Breast Neoplasms; Carcinoma, Ductal, Breast; Cell Nucleus - ultrastructure; Cervix Uteri - cytology; Cluster Analysis; cluster validation; Clustering algorithms; Discriminant Analysis; Female; Gaussian mixture model; Gaussian processes; Histocytochemistry - methods; Hospitals; Humans; Image analysis; Image Processing, Computer-Assisted - methods; Image segmentation; Investigative techniques, diagnostic techniques (general aspects); Medical diagnostic imaging; Medical sciences; Miscellaneous. Technology; Normal Distribution; overlapped nuclei segmentation; Pathology; Pathology. Cytology. Biochemistry. Spectrometry. Miscellaneous investigative techniques; Reproducibility of Results; Surface topography; unsupervised Bayesian classifier; Segmentation; Image processing; Cell nucleus; Automatic cell segmentation; cluster validation; Mixture theory; overlapped nuclei segmentation; Overlap; Unsupervised classification; Gaussian mixture model; Gaussian process; Microscopy; unsupervised Bayesian classifier; Cell; Biomedical engineering
• ISSN: 0018-9294; 1558-2531
• DOI: 10.1109/TBME.2010.2060486

In a fully automatic cell extraction process, one of the main issues to overcome is the problem related to extracting overlapped nuclei since such nuclei will often affect the quantitative analysis of cell images. In this paper, we present an unsupervised Bayesian classification scheme for separating overlapped nuclei. The proposed approach first involves applying the distance transform to overlapped nuclei. The topographic surface generated by distance transform is viewed as a mixture of Gaussians in the proposed algorithm. In order to learn the distribution of the topographic surface, the parametric expectation-maximization (EM) algorithm is employed. Cluster validation is performed to determine how many nuclei are overlapped. Our segmentation approach incorporates a priori knowledge about the regular shape of clumped nuclei to yield more accurate segmentation results. Experimental results show that the proposed method yields superior segmentation performance, compared to those produced by conventional schemes.