PhysioNet Challenge Nets Success
“That’s a wrap!”
One Heart Health had the privilege to participate in the 2022 PhysioNet Challenge. This international challenge draws participants from both academic institutions and industry experts. This year teams worked on pediatric heart sounds collected from two mass screening campaigns conducted in Northeast Brazil.
Our algorithm team, led by Dr. Matheus Araujo, a native of Brazil, with contributions from Dewen Zeng, Xirong Xi, Dr.Yiyu Shi (University of Norte Dame), Dr. Joao Palotti (Qatar Computing Research Institute), Dr. Lee Pyles (West Virginia University) and Dr. Quan Ni (One Heart Health), investigated several deep learning strategies to improve their performance to identify abnormal murmurs.
Overall, we placed 17th, but our solution was only 5% away from 1st Place! Because the challenge is open source, we also learned how to implement the best solution. We will implement learnings from this competition directly into our development of automatic detection of congenital heart diseases using the One Heart app.
Read more about our solution and results published here:
- Maiby’s Algorithm: A Two-stage Deep Learning approach for Murmur Detection in Mel Spectrograms for Automatic Auscultation of Congenital Heart Disease
This technology will increase our ability to scale CHD Screening across the world in low-resource areas, but we won’t be able to do that on our own. Please consider donating to One Heart Health and support us as we deliver on our vision that: Every child with heart disease gets a chance to live a healthy life. We’ve transformed over 26,000 children’s lives through One Heart Health’s heart disease screening programs. You can help transform thousands more. Click here to donate or contact give@onehearthealth.org to learn more.
