Automated segmentation and classification of unlabelled malaria parasites in red blood cells from phase contrast images

An automated method for detection of unlabelled malaria parasites in red blood cells is presented. From phase-constrast microscopy images of live red blood cells, the proposed algorithm segments red blood cells and classifies them into uninfected and infected cells. This approach can be used to o...

Full description

Saved in:
Bibliographic Details
Main Author: Mathew Athul Mangalathumannil
Other Authors: Justin Dauwels
Format: Theses and Dissertations
Language:English
Published: 2016
Subjects:
Online Access:http://hdl.handle.net/10356/68682
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-68682
record_format dspace
spelling sg-ntu-dr.10356-686822023-07-04T15:04:53Z Automated segmentation and classification of unlabelled malaria parasites in red blood cells from phase contrast images Mathew Athul Mangalathumannil Justin Dauwels School of Electrical and Electronic Engineering DRNTU::Engineering::Electrical and electronic engineering An automated method for detection of unlabelled malaria parasites in red blood cells is presented. From phase-constrast microscopy images of live red blood cells, the proposed algorithm segments red blood cells and classifies them into uninfected and infected cells. This approach can be used to observe the progression of malaria parasites through different stages as the algorithm presented can be used on unstained images. In related work, classification is performed on stained images. The image processing framework consists of image pre-processing, segmentation and classification of infected and uninfected cells from phase contrast images. Two segmentation methods; morphological operations and template matching, are applied and compared. To separate connected cells, a watershed algorithm is applied. The final goal of this project is to detect malaria infected cells from microscopic blood images at lowest magnification and then control the imaging system in the microscope to automatically zoom to monitor the infected cells. Manual detection of malaria from a cell sample is a very tedious and time consuming process. Using our methodology we are able to differentiate infected and non-infected cells automatically. Our approach also bypasses the need to stain the cell samples. This work can thus be implemented for fast, simple to execute and efficient diagnosis of malaria infected cells. Master of Science (Computer Control and Automation) 2016-05-30T08:52:30Z 2016-05-30T08:52:30Z 2016 Thesis http://hdl.handle.net/10356/68682 en 63 p. application/pdf
institution Nanyang Technological University
building NTU Library
continent Asia
country Singapore
Singapore
content_provider NTU Library
collection DR-NTU
language English
topic DRNTU::Engineering::Electrical and electronic engineering
spellingShingle DRNTU::Engineering::Electrical and electronic engineering
Mathew Athul Mangalathumannil
Automated segmentation and classification of unlabelled malaria parasites in red blood cells from phase contrast images
description An automated method for detection of unlabelled malaria parasites in red blood cells is presented. From phase-constrast microscopy images of live red blood cells, the proposed algorithm segments red blood cells and classifies them into uninfected and infected cells. This approach can be used to observe the progression of malaria parasites through different stages as the algorithm presented can be used on unstained images. In related work, classification is performed on stained images. The image processing framework consists of image pre-processing, segmentation and classification of infected and uninfected cells from phase contrast images. Two segmentation methods; morphological operations and template matching, are applied and compared. To separate connected cells, a watershed algorithm is applied. The final goal of this project is to detect malaria infected cells from microscopic blood images at lowest magnification and then control the imaging system in the microscope to automatically zoom to monitor the infected cells. Manual detection of malaria from a cell sample is a very tedious and time consuming process. Using our methodology we are able to differentiate infected and non-infected cells automatically. Our approach also bypasses the need to stain the cell samples. This work can thus be implemented for fast, simple to execute and efficient diagnosis of malaria infected cells.
author2 Justin Dauwels
author_facet Justin Dauwels
Mathew Athul Mangalathumannil
format Theses and Dissertations
author Mathew Athul Mangalathumannil
author_sort Mathew Athul Mangalathumannil
title Automated segmentation and classification of unlabelled malaria parasites in red blood cells from phase contrast images
title_short Automated segmentation and classification of unlabelled malaria parasites in red blood cells from phase contrast images
title_full Automated segmentation and classification of unlabelled malaria parasites in red blood cells from phase contrast images
title_fullStr Automated segmentation and classification of unlabelled malaria parasites in red blood cells from phase contrast images
title_full_unstemmed Automated segmentation and classification of unlabelled malaria parasites in red blood cells from phase contrast images
title_sort automated segmentation and classification of unlabelled malaria parasites in red blood cells from phase contrast images
publishDate 2016
url http://hdl.handle.net/10356/68682
_version_ 1772827490875604992