Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

A review of automatic malaria parasites detection and segmentation in microscopic images

 
: Rosado, Luis; Correia da Costa, José M.; Elias, Dirk; Cardoso, Jaime S.

Anti-infective agents 14 (2016), Nr.1, S.11-22
ISSN: 2211-3525 (print)
ISSN: 2211-3533 (Online)
Englisch
Zeitschriftenaufsatz
Fraunhofer AICOS ()
supervised classification; Computer Aided Diagnosis; feature extraction; image analysis; malaria; segmentation

Abstract
Background: Malaria is a leading cause of death and disease in many developing countries, where young children and pregnant women are the most affected groups. In 2012, there were an estimated 207 million cases of malaria, which caused approximately 627 000 malaria deaths. Around 80% of malaria cases occur in Africa, where the lack of access to malaria diagnosis is largely due to a shortage of expertise, being the shortage of equipment the secondary factor. This lack of expertise for malaria diagnosis frequently results on the increase of false positives, since prescription of medication is based only on symptoms. Thus, there is an urgent need of new tools that can facilitate the rapid and easy diagnosis of malaria, especially in areas with limited access to quality healthcare services.
Methods: Various image processing and analysis approaches already proposed on the literature for the detection and segmentation of malaria parasites in blood smear microscopic images were collected and reviewed. This timely review aims to support the increasing interest in the development of low cost tools that can facilitate the rapid and easy diagnosis of malaria, especially in areas with limited access to quality healthcare services.
Results: Malaria parasites detection and segmentation techniques in microscopic images are, in general, still in need of improvement and further testing. Most of the methodologies reviewed in this work were tested with a limited number of images, and more studies with significantly larger datasets for the evaluation of the proposed approaches are needed. Despite promising results reported during the past years, the great majority of the computer-aided methods found on the literature for malaria diagnosis are based on images acquired under well controlled conditions and with proper microscopic equipment. However, one should take into account that 80% of malaria cases occur in Africa, where this type of equipment is scarce or even nonexistent in common healthcare facilities.
Conclusion: This work collects and reviews various image processing and analysis approaches already proposed on the literature for the detection and segmentation of malaria parasites in blood smear microscopic images. This timely review aims to support the increasing interest in the development of image processing-based systems to be used in rural areas of developing countries, which might be the next future trend in malaria computer-aided diagnosis.

: http://publica.fraunhofer.de/dokumente/N-421963.html