Hep-2 cell images fluorescence intensity classification to determine positivity based on neural network

This paper applies the concept of Artificial Neural Network (ANN) to classify fluorescence intensity of Hep-2 cell images into three classes; positive, intermediate and negative auto-immune disease. Recently, the recommended method for detection antinuclear auto-antibodies (ANA) is Indirect Immunofl...

Full description

Main Authors: Zazilah, M., Mansor, A.F., Yahaya, N.Z.
Format: Conference or Workshop Item
Institution: Universiti Teknologi Petronas
Record Id / ISBN-0: utp-eprints.31554 /
Published: Institute of Electrical and Electronics Engineers Inc. 2015
Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-84946595063&doi=10.1109%2fISTT.2014.7238192&partnerID=40&md5=a9855bb7805e2cf23b6736f1ea0b4d3f
http://eprints.utp.edu.my/31554/
Tags: Add Tag
No Tags, Be the first to tag this record!
Summary: This paper applies the concept of Artificial Neural Network (ANN) to classify fluorescence intensity of Hep-2 cell images into three classes; positive, intermediate and negative auto-immune disease. Recently, the recommended method for detection antinuclear auto-antibodies (ANA) is Indirect Immunofluorescence (IIF). The diagnosis consists of estimating fluorescence intensity in the cells. Since the increasing of test demands, trained personnel are not always available for these tasks and the identification of positivity has recently done manually by human analyzing the slide with a microscope, leading to subjective and bad quality results. This work will develop Computer Aided Diagnosis (CAD) tools that can offer a support to physician decision. Then, it discusses image preprocessing, image segmentation and feature extraction. Later, this lead to the proposal of ANN-based classifier that is able to separate essentially the intermediate sample of ANA diseases. The approach has been evaluated using 142 cell images, for 372 training data. The measured performance shows a low overall error rate which is 3 , this is lower than error rate of observed intra-laboratory variability. © 2014 IEEE.

Similar Items