Radial Tracing Method of Cytoplasm Segmentation in Overlapped Cervical Cell Images
M. Mohideen Fatima alias Niraimathi, V. Seenivasagam · IETE Journal of Research · 2015
This paper proposes a new method to trace the obscure boundary of the cytoplasm in stained microscopic cervical cell images because of low intensity contrast with the background and the contaminations caused by inflammatory cells, blood stains, etc. A Gaussian filter is used for smoothing the image for suppressing the high frequencies. A bi-group enhancer is used to suppress the noise and to improve the contrast between the weak boundaries of the objects of nuclei and cytoplasm of cervical cells. The binarized bi-group image is used for nuclei segmentation. But the segmentation of cytoplasm in overlapped cervical cell images is really a very tedious task because of the fuzzy boundaries existing between cell images and the vague boundary which separates the cytoplasm and background. Hence, a gradient map is generated which is used as a base to radial trace the boundary of the cytoplasm to locate ambiguous boundaries. Once the true boundary of nuclei and cytoplasm is located, the quantitative metrics (e.g. nuclei area, cytoplasm area, the nucleus–cytoplasm ratio, etc.) can be calculated since cells in cancers and pre-cancers are characterized by many morphologic and architectural alterations, including shape and size of the cytoplasm and nucleus, an increasing nuclear–cytoplasmic ratio, etc.