Automated malaria parasite detection using a high tolerance image analysis technique

Abstract

In this paper a robust method for automated malaria diagnosis is presented. The robustness of the method is due to the addition of the image preclassification step which categorizes the image into one of the determined classes. The developed technique was able to detect positive cases correctly with an accuracy of 89% and negative cases with 91% accuracy. The technique's specificity and sensitivity were observed to be at 77% and 96% respectively.