Real-Time ECG Signal Feature Extraction for the Proposition of Abnormal Beat Detection - Periodical Signal Extraction

This paper presents a novel algorithm for ECG signal feature extraction and proposes a solution for detection of abnormal beats. The main goal is to extract signal and heartbeat parameters by using local extreme values and their dependences. This produces the averaged heartbeat which is used to analyze the signal in real time to detect anomalies. The main advantages of our approach are fast computation and portability between software development platforms. The main disadvantage is the need of clear fragments of samples for the training of the algorithm. Experimental results indicate the proposed system is efficient when compared to other models.