Morphological segmentation of clustered nuclei in analytical cytology

Cluster division is a crucial problem in analytical image based cytology. Existing algorithms often make use either of the image content or its geometric characterization. Both sources of information are needed for a correct segmentation of difficult clusters. Morphological algorithms naturally deal...

Full description

Saved in:
Bibliographic Details
Main Authors: Malpica, N., Santos, A., de Solorzano, C.O., Vaquero, J.J., del Pozo, F., Garcia-Sagredo, J.M.
Format: Conference Proceeding
Language:English
Subjects:
Blood
Cameras
Cells (biology)
Charge coupled devices
Clustering algorithms
Fluorescence
Image analysis
Image segmentation
Marine animals
Microscopy
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Cluster division is a crucial problem in analytical image based cytology. Existing algorithms often make use either of the image content or its geometric characterization. Both sources of information are needed for a correct segmentation of difficult clusters. Morphological algorithms naturally deal with the object oriented criteria such as shape, size, contrast, connectivity, etc. In this paper, we present and evaluate a morphological watershed based algorithm applied to fluorescence stained clustered nuclei division. Results are shown for two different types of samples namely bone marrow and peripheral blood specimens. These results are better than those obtained for other published algorithms. This algorithm can also be easily adapted to different types of specimens.
DOI:10.1109/IEMBS.1996.651929