Cytological image contour extraction and region segmentation.
Item
-
Title
-
Cytological image contour extraction and region segmentation.
-
Identifier
-
AAI9020777
-
identifier
-
9020777
-
Creator
-
Li, Yuan.
-
Contributor
-
Adviser: Joseph Barba
-
Date
-
1990
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Engineering, Electronics and Electrical | Engineering, Biomedical
-
Abstract
-
In this doctoral dissertation, multi-thresholding combined with binary morphological edge detection and region extraction approaches for cytological images are investigated. Specifically, techniques for detection of the cell and nuclear contours in cytological specimens are presented. Gray level images are viewed as stacked binary bit plane with each bit plane image representing the effect of thresholding the gray level image with multiple gray level windows. A binary morphological edge detector is employed to identify edges on selected bit planes. A simple contour tracking algorithm incorporating a priori information of the cell structure is used to identify the most likely cell and nuclear contours. Further refinery is made by matching each binary edge with the gradient of the original image. A figure of merit is defined and used for selecting specific binary edges (root edges) which most likely contain the contours of the cell image with best extractability. A fast and effective technique is also presented for determining root edges automatically based on searching binary edges in the spatially shifted images. Additionally, a two-step hierarchical region and texture segmentation technique by using vector quantization is presented. The regions encoded by low rate vector quantizers are refined by searching the replacement of a much larger codebook containing all the edges information. Cell regions are also segmented into texton regions. Related texture features are then extracted for further analysis and classification.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
