A novel method for EKG anomaly detection based on the double sliding window technique

Traditional detection methods often rely on fixed thresholding or machine-learning models, which can be computationally expensive. This study introduces a double-sliding-window technique to detect anomalies in ECG signals through adaptive-signal analysis. This method uses two independent sliding windows to dynamically track signal variations, enabling real-time identification of deviations in heart rhythm without prior knowledge of data distribution. The detection module achieved an average execution-time of 0.03 seconds with using less than 16 MB of memory, supporting its suitability for real-time systems. Furthermore, confusion matrix analysis yielded an accuracy of 95.33%, a sensitivity of 95.00%, a precision of 100.00%, and a specificity of 100.00%.