PhysioNet challenge 2020

A good result in a Cardiology Challenge

Konstanz, October 2020. Lucas Weber successfully participated in this year's Physionet/Computing in Cardiology Challenge (2020). 

 Rastergrafik

This challenge is a major event in the world of biosignal analysis. Every year, one problem from the medical field is selected for this challenge. The main motivation is to develop validated open-source algorithms for important medical problems to support physicians in their daily work. This year's challenge focused on the automatic detection of cardiac disorders detected by 12-lead ECGs. Heart disorders are the leading cause of death worldwide. Continuous signal monitoring is resource-intensive, automated detection of abnormalities offers many advantages and can, among other things, reduce the probability of false diagnoses or avoid delayed detection. This year's challenge consisted of over 43,000 ECG signals. The heterogeneity is a challenge for most algorithms but also offers a large variety of data for machine learning. Teams from all over the world participated in the challenge. The Ubiquitous Computing Lab, under the direction of Prof. Dr. Seepold, submitted a contribution based on Deep Learning, which offers the possibility to learn from and process signals of different lengths. Out of 163 teams that qualified for the challenge, 70 submitted a successful contribution for the covert test set, of which 39 teams met all criteria (reproducible results, open-source license, and appropriate presentation of the approach). We are proud to have reached the fifteenth place with our idea. Information about the challenge and the ranking can be found at https://physionetchallenges.github.io/ . The project was supported by a small research project of HTWG.

The link to the HTWG article is this.


News Archive