Image Analysis of Blood Slides for Automatic Malaria Diagnosis

Poostchi, Mahdieh ;   Ersoy, Ilker ;   Bansal, Abhisheka ;   Palaniappan, Kannappan ;   Antani, Sameer ;   Jaeger, Stefan ;   Thoma, George

Malaria is a serious global health problem, claiming the lives of 450,000 children per year. A fast and reliable test for diagnosing malaria would be a promising approach to fight this disease. We present an automatic system for diagnosing and quantifying a malaria infection in cultured red blood cells on thin films, using image processing techniques. We measure an average error of 1.8% by comparing the true frequency of infected cells with the automatically computed infection frequency, which encourages applying our technique for malaria diagnosis in the field.