Community Research and Development Information Service - CORDIS

ECG automatic analysis toolkit

HEARTS adopted two different ways to focus and analyse heart disease diagnosis: the disease-centric and the patient-centric approaches.

Based on the concept of “adaptive behavioural analysis”, the decision support module analyses data through innovative neuro/fuzzy processing technologies, obtaining anytime information about health status related to the specific subject and the specific context.

The main objective was to develop techniques for early detection of myocardial ischaemia starting from the ECG signal. This result has been achieved through the integration of the disease-centric and the patient centric approach, trying to give immediate warning to the users when there are signs of myocardial ischaemia, together with easy interpretability and reasoning capabilities.

The architectural structure of the decision support subsystem is modular, hierarchical and aimed to make the process understandable to the medical experts. The so implemented components have much potential in the hospital or medical school settings.

The patient-centric approach is based on the ECG beat recognition method and it’s based on a supervised neural network-based algorithm.

Several experiments have been deployed both on data gathered during the validation and test phases and on the European ECG database.

The results of experiments have shown the good efficiency of this second solution. Indeed, with the novel approach, the accuracy of beat classification for several patient records taken from European ECG database was over 94%. The investigation shows that the proposed beat classifier is very reliable, and hence it may be a useful practical tool for automatic detection of ischemic episodes.

The added value given by the integration of the above approaches is furthermore significant, making the final integrated decision support module able to be very high for both sensitivity and specificity in the identification of ischemic beats (ST segment deviation) for the specific patient.

The decision support modules developed by HEARTS can be used not only as part of the complete telemedicine system but also as a standalone module in the hospital or embedded in medical devices. They can also be integrated to other decision support systems in the hospital, especially in the operating theatre intensive care units and coronary care units where the patient’s ECG and other biomedical data are continuously monitored and the early detection of myocardial ischaemia is important.

Good results have been achieved both on European ST-T database segments and on ECG segments collected during verification and validation phases.

More information on the Hearts project can be found at:

Related information

Reported by

University of Sunderland
Langham Tower Ryhope Road
SR2 7EE Sunderland
United Kingdom
See on map
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top