Malaria Diagnostics Software
The new diagnostics approach for malaria detection works with a mobile phone and a microscope. At least 15 images of a thick blood smear of a single patient are required to test for malaria. The software can distinguish between plasmodium falciparum and plasmodium vivax. Per test four to five neural networks are required. The computation is then performed in the cloud and will take under 3 minutes. The advantages are the consistent performance over time as well as the speed and accuracy. Connecting this software into a patient management system would further reduce errors and increase the time from the test to the result.
Next Steps
In order to acquire a medical certification, a non-inferiority study is required to compare the new malaria diagnostics software to the standard microscopy procedure. For this, as well as for the medical certification itself, the team is looking for funding.