Despite the global effort to fight malaria transmission, nearly 250 million cases of malaria were reported in 85 countries in 2020. Sadly, the Global Health Observatory reports that even with the newest treatments, mortality increased by 12%. Now, medical researchers are hoping to use embedded vision to help diagnose malaria faster and get people treated before cases become severe.
How Microscopy Is Used to Identify Malaria Infections
To diagnose malaria cases, doctors currently use a microscope with a very limited field of view. It can take up to 15 minutes per slide to make a proper diagnosis. With soaring infection rates and limited physicians, some areas don’t have the doctors needed to scan each sample. Microscopy (the use of microscopes to view objects that can’t be seen with the naked eye) traditionally uses a single source and type of visible light. But a new method for detecting malaria uses a different approach.
Engineers designed a microscope with a new lighting pattern that can help quickly detect red blood cells infected with the malaria parasite. A bowl-shaped LED light source illuminates samples from different angles with different colors of light. Shadows and highlights accentuate different features of the blood sample so that artificial intelligence can be used to find malaria faster and with higher accuracy.
Training AI to Find Evidence of Malaria
Physicians must often look through a thousand cells to find a single malaria parasite. Because of the magnification required, they can only examine about a dozen at a time. Well-trained physicians can typically identify malaria cases accurately with about a 75% success rate. But researchers were able to build a system that correctly classifies malaria over 90% of the time.
How did they do it? By feeding the microscope hundreds of samples of malaria-infected red blood cells, the microscope equipped with embedded vision was able to learn the most important features for diagnosing malaria. And the microscope can identify those features within a wider field of view, making the diagnosis in seconds.
Researchers can see much promise in this technique. Examining slide after slide can lead to fatigue for humans. But a microscope with embedded vision never gets tired of scanning slides. Physicians hope to see this technology expanded to find other types of infections and help fill the widening gap between the number of physicians available and those needed.
Looking for a quality camera for your facility? Browse the Phase 1 catalog to find the best camera for your next embedded vision project.