Interactive Deep Learning for ECG Monitoring

Case ID:

This invention is a deep learning algorithm system that can incorporate human-interpretable explanations in its training process for the analysis of ECG signals. This method takes into account classification loss, the influence of unimportant features to model prediction, the influence of important features to model prediction, and user feedback. This invention may use real-time ECG signals to provide clinical decision support to health care providers in the form of a web or mobile application.  This technology can enable a faster and more reliable method of reading ECGs and help screen and identify important image features that require expert scrutiny.

Deep learning has gained significant attention as a subset of machine learning with significant investments in research over the last few years. Deep learning is an advanced type of artificial neural networks. This technology largely emulates the human brain by learning and solving complex problems, which may be applied to the interpretation of ECG signals where resources in trained ECG experts and/or health care reimbursement are limited. Deep learning is currently the fastest growing segment of artificial intelligence due to its ability to extract information from unstructured data.


  • ECG image classification
  • Medical imaging classification


  • Capable of multiple classifications
  • Increased accuracy and sensitivity
  • Deep learning improved with human feedback
Patent Information:
Contact For More Information:
Jonathan Larson
Senior Licensing Manager, College of Science
The University of Arizona
Lead Inventor(s):
Chicheng Zhang
Dharma KC
Christopher Gniady
Parth Agarwal