Students in the biomedical engineering class developed a web and mobile application that uses deep learning to diagnose measles through analysis of skin lesions that provides results in 0.25 seconds, unlike PCR and other tests. immunoglobin which take longer.
William Chonzie, Biomedical Engineering student leader, said: “Transfer learning techniques were used using Keras and TensorFlow. A larger dataset was curated by combining images from three different sources.
Guided by their professors, the students applied different scientific processes and procedures to improve the model’s performance.
Measles remains one of the leading killer diseases in Malawi, particularly among infants. As an airborne disease, it spreads quickly. Student innovation can ensure that communities can detect such cases at an early stage and seek medical attention.
Manufacturing engineering professor Peter Mwambananji pioneered the invention of a system that uses machine learning to analyze soil and environmental factors to recommend crops suitable for agriculture. Now the system relies only on static weather data. It intends to use AI to further train the system to retrieve dynamic data from online sources via API calls.
“As the world is go digitaljust like most resources. Having power at your fingertips is very important. Farmers who use crop recommendation systems are assured of quick results based on data-driven evidence,” says Mwambananji, adding that “the service is inexpensive because it will only use the internet if they prefer the web app or smart mobile app. platforms, and also free for the USSD version.
Similarly, another MUST student, Syton Mphaka, who is pursuing a bachelor’s degree in meteorology and climate science, has developed an early detection system for droughts and floods.
Mphaka’s innovative system detects water levels and processes the data. It works using the phone’s SIM card and is therefore capable of sending messages and activating an alarm or sirens as a warning.
When the system detects high water levels in rivers, it automatically sends messages to communities within a 300 meter radius and activates sirens to warn them of the danger.
“The system has sensors that are used to collect data on water level changes. Changes in water levels are the main indicator of flooding or drought. The display will be installed in the community where people will go and see the water levels without necessarily going to the rivers around them because the process will be done digitally,” says Mphaka.
Innovations such as those developed by MUST students and faculty demonstrate the value that academic institutions have in helping to solve pressing community challenges. The Malawi government must support such efforts that harness artificial intelligence to make citizens’ daily lives easier.
Benson Kunchezera is a Malawian independent journalist based in Blantyre.
bkunchezera84@gmail.com
