MALARIA REKOGNIZER is an application for Windows (8 and 10) based on Artificial Intelligence that helps considerably to reduce times, costs, and error margin in a thin blood smear. It replaces the manual counting of parasites present in cells.



Manual count (human)
Malaria Rekognizer: Remarkable faster

Remarkable reduction of error margin

Effectiveness Percentage > than 96% (see image 1)
Margin of error < than 4%

Image 1: Effectiveness Percentage

How does it work?

In a thin blood smear, the laboratory professional must capture microscopic images with a 100X magnification

You can use a mobile or camera to take the photos (jpg or png formats)
It’s recommended that there are a maximum of 20 images of fields with 250 cells each one of them

The captured microscopic fields (images) are attached to Malaria Rekognizer

Parasite counting isn’t necessary

The obtained results are:

Total number of infected samples
Total number of NOT infected samples
Infected percentage
NOT infected percentage
Automatic segmentation of microscopic fields in images (by default 130 px) of cells or parts of them
Automatic generation of .csv file with detailed info of each analyzed cell (name, status and infection percentage)

According to the number of infected and uninfected cells, the percentage of parasitic density is deduced, and it’s determined if the patient has malaria (see image 2)

Image 2: Malaria Reckognizer Interface

Used technologies

Artificial Intelligence