A Nigerian Artificial Intelligence (AI) startup company Ubenwa took on an important challenge, saving infants from asphyxia. Birth asphyxia is one of the top three causes of infant mortality in the world. Ubenwa is an Igbo word which means the cry of a child.
The AI company which is composed of Nigerians developed a machine learning system that can take as input the infant cry, analyse the amplitude and frequency patterns in the cry, to provide instant diagnosis of birth asphyxia.
According to reports, 95% accuracy has been achieved in trials involving over 1,400 pre-recorded baby cries. The United Nations stated, “If newborns who have asphyxia can be detected early enough, we may be able to save their lives.”
Read: Cameroonian innovator Arthur Zang wins Engineering Innovation Prize for Cardio-Pad
https://twitter.com/HelloDanio/status/939262192192446464
Charles Onu, the principal innovator of the startup said “Ubenwa, deployed via mobile devices and wearables, will reduce the skill and cost required to diagnose the condition as well as make swift detection possible.” The company applied for the $5million IBM Watson XPrize.
900,000 babies die every year from perinatal asphyxia. The detection of baby birth asphyxia takes 10 seconds with the Ubenwa application. It is faster compared to blood test based detection. Compared to the current method using a blood gas analyser, Ubenwa is non-invasive (requiring only cry rather than blood), low-cost (about 95% cheaper than clinical alternative), and requires little or no skill to operate.
Read: Top 10 nominees for Innovation Prize for Africa announced
According to F65, Ubenwa will help parents to monitor and better care for their children, while also providing health institutions (especially in developing countries) with an effective, low-cost alternative.
Onu is supported by Dr. Eyenimi Ndiomu who serves as the clinical leader for the team. Innocent Udeogu is the software engineering lead while Urbain Kengni is the business development lead. Onu started Ubenwa Intelligence Solutions in June 2014.
