Development of an Auto-detection and Quantification Algorithm Of Malaria Infection Using Image Processing

Most of the malarial diagnostic methods either depend on manual counting of infected red blood cells or requires complex laboratory facilities. In both cases, the diagnostic is time-consuming, expensive, requires trained personnel, sometimes produce erroneous results due to manual intervention, and hinders rapid diagnostics of malarial infection. Malaria is mostly fatal if not diagnosed and treated promptly, therefore, it is imperative to devise a methodology that provides a rapid, cost-effective, and accurate malarial diagnosis with proper quantification. Here, we propose an image processing-based malaria detection methodology using support vector machine (SVM) that can detect and quantify malarial infection with up to 96% accuracy. The image processing algorithm is implemented on a range of images and the outcomes are in good agreement with the actual diagnostic results thereby, validating the methodology.