Disclosed herein are methods and apparatus for making a determination whether a blood sample is, or is not, infected with a
malaria parasite. The determination is made using a trained
machine-learning (ML)
algorithm. In some aspects, a
microfluidic chip is used for the concentration of red blood cells infected with these parasites from uninfected red blood cells, the
staining of the blood sample which differentially stains infected and uninfected red blood cells, and holds the sample of interest for imaging. The
microfluidic chip is inserted into an optical subsystem that magnifies the image created from
transmitted light microscopy. The magnified image is captured with a camera, and a trained ML
algorithm assesses if the sample does or does not contain the parasite. Generally, the ML
algorithm is executed on a portable computing device such as a smartphone.