Analysis of Cell Images for Identification of Cancer.

Date of Submission

December 1999

Date of Award

Winter 12-12-2000

Institute Name (Publisher)

Indian Statistical Institute

Document Type

Master's Dissertation

Degree Name

Master of Technology

Subject Name

Computer Science

Department

Machine Intelligence Unit (MIU-Kolkata)

Supervisor

Mandal, Deba Prasad (MIU-Kolkata; ISI)

Abstract (Summary of the Work)

Cervical cancer or cancer of the cervix, a common kind of cancer in women, is a disease in which cancer (malignant) cells are found in the tissues of the cervix. It presents a significance challenge to the bealth care community as it is the second most common cause of cancer death among women world-wide (1, 2].The present investigation is an attempt towards finding suitable nethodologies to identify pre-cancerous changes in the tissues of cervix based on pattern recogni- tion and image processing tools like thresholding, fuzzy c-means cluster seeking algorithm, mathematical morphology, textural feature etc.1.1 Motivation Cancer of the cervix usually grows slowly over a period of time. Before cancer cells are found on the cervix, the tissues of the cervix go through changes in which cells that are not normal begin to appear (known as pre-cancerous cella). Later, cancer cells start to grow and spread more deeply into the cervix and to sutrounding areas.Pre-cancerous and cancerous changes of the cervix can easily be found by pap test (also known as pap smear). This test involves scraping some cells from the surface of the cervix, smearing them on to glass slides and analyzing the slide images by looking the slides under a microscope.Every alide image has two membranes (lines) known as basal merhbrane and su-perficial membrane respectively. The portion between the basal and superficial mbranes is only analyzed in the pap test. This portiou has four different regions Bamely, i) basal region which lies just aboye the basal membrane, ii) parabasal region which line above the basal region, ii) intermediate region which lies above the parabasal rngion and iv) superficial region which lies between the interme- diate region and superficial menmbrane. Depending on availability of the imma- ture/irregular cells in different regions, the images are characterized either as nor- mal or cervical intraepithelial neoplasias (CINS) (i.e., CIN-1 (mild displasia), CIN-2 (moderate displasia) or CIN-3 (severe displasia)).1.2 Scope of the Thesis The present investigation is besically concerned with finding automatic methodolo- gies to identify pre-cancerous changes in the tissues of the cervix as done in pap test. Initially a cooccurrence matrix (of gray values) based entropic thresholding method is applied on the slide images where mostly all the cell nuclei along with few other particles (lying in the stroma) are found to get segmented. The overlap- ping nuclei are then decomposed using a newly proposed algorithm. Each image is now rotated around its center (decided based on the nearly hamogeneous region above the superficial membrane) to make superficial membrane (as well as basal membrane) horizontal and top side in the image. Using texture features promi- nence and shading on rectangular windows around every objeat pixel, the fuzzy C-means algorithm is found to extract out the homogeneous aren above the su- perficial line and accordingly the superficial line is identified. Depending on the orientation of the (cell) nuclei, the basal line is found. The portion between basalmembrane) horizontal and top side in the image. Using texture features promi- nence and shading on rectangular windows around every objeat pixel, the fuzzy c-means algorithm is found to extract out the homogeneous area above the su- perficial line and accordingly the superficial line is identified.

Comments

ProQuest Collection ID: http://gateway.proquest.com/openurl?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&res_dat=xri:pqm&rft_dat=xri:pqdiss:28843239

Control Number

ISI-DISS-1999-70

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

DOI

http://dspace.isical.ac.in:8080/jspui/handle/10263/6242

This document is currently not available here.

Share

COinS