Enhanced QRS detection and ECG compression using adaptive thresholding: A real-time approach for improved monitoring and diagnosis

• QRS detection using threshold algorithms combined with absolute value curve length transform (A-CLT) as a novel approach. • Adaptive thresholds achieve 99.69 % sensitivity, 99.60 % predictability, and 90.63 % F1 score from the MIT-BIH arrhythmia database. • Lossless compression technique using first derivative and entropy coding. • High speed and small space occupation for portable and wearable ECG. Electrocardiograms (ECGs) heavily rely on QRS detection; however, most existing R-peak detectors face challenges due to varying QRS morphology and non-stationary signals. To address these issues, an enhanced real-time method for detecting QRS complexes and compressing ECG signals is proposed, particularly beneficial for Internet of Things (IoT) applications in biomedicine. This method integrates advanced signal processing algorithms with an adaptive threshold technique to improve accuracy and efficiency. By dynamically adjusting detection thresholds based on local ECG signal characteristics and incorporating wavelet transform and morphological operations, robust QRS detection is achieved, even in noisy environments. The suggested QRS detector has been verified using 48 ECG records from the MIT-BIH arrhythmia database and 15 ECG records from the BIDMC Congestive Heart Failure Database (CHF). Mean sensitivity, positive predictivity, and F1 score of 99.69 %, 99.60 %, and 99.63 %, respectively, are achieved for the MIT-BIH arrhythmia database, and 99.56 %, 99.05 %, and 99.29 %, respectively, for the BIDMC CHF database. Also, a lossless compression technique, utilizing the ECG signal's first derivative and entropy encoding, is incorporated into the architecture, eliminating complex data processing and storage requirements. Overall, this approach offers an efficient, cost-effective solution for medical signal monitoring and diagnosis on low-power hardware suitable for mobile applications.