Title

An automated ovarian tissue detection approach using type P63 non-counter stained images to minimize pathology experts observation variability

Document Type

Conference Proceeding

Publisher

IEEE

School

School of Science

RAS ID

23249

Comments

Originally published as: Sazzad, T. S., Armstrong, L. J., & Tripathy, A. K. (2016, December). An automated ovarian tissue detection approach using type P63 non-counter stained images to minimize pathology experts observation variability. In Biomedical Engineering and Sciences (IECBES), 2016 IEEE EMBS Conference on (pp. 155-159). IEEE. Original article available here

Abstract

Ovarian conceiving tissues are essential for future generation. Women may face conceiving complications at any age and need to consult with medical experts for proper treatment. It is required to know the condition of the ovary for which ovarian biopsy test is required in the histopathology laboratory. Histopathology experts consider manual microscopic analysis as a viable approach in compare to electronic scanners. The main drawback of manual analysis approach is observation variability. Computer based approaches have become popular due to the fact that it minimizes observation variability, same results can be achievable, possible to minimize time and accuracy rate can be improved in compare to manual analysis. To date research works for human ovarian tissues are limited in compare to cancer cells using especially 100x magnification images. Experts use different magnifications if they are unable to analyze any region accurately. None of the existing computer based approaches have considered different magnifications for the same regions for analysis. In this paper a new modified fully automated approach has been proposed which is faster than all existing approaches where 3 different magnifications were incorporated to identify the same region with different magnifications to minimize the expert variability issues. The proposed approach was able to increase the accuracy rate while minimizing the observation variability for manual analysis.

DOI

10.1109/IECBES.2016.7843434

Access Rights

Not open access

Share

 
COinS