ECG Signal Processing & Waveform Telemetry Suite
Automated Pan-Tompkins QRS detection and clinical cardiac telemetry feature extraction.
A biomedical signal-processing pipeline engineered for Sagar's B.Tech final-year capstone. Implements dual-pass bandpass Butterworth filtering, 5-point derivative filtering, non-linear squaring, and moving-window integration to extract Pan-Tompkins QRS complexes and Heart Rate Variability (HRV) metrics from MIT-BIH arrhythmia streams in real-time with sub-2ms processing latency.
The Problem
Raw electrocardiogram electrode telemetry suffers from severe muscle artifact noise, baseline wander, and electrode displacement, making reliable automated cardiac anomaly triage difficult.
The Approach & Implementation
Implemented a classic Pan-Tompkins pipeline in Python, NumPy, and SciPy: 5–15 Hz bandpass digital filtering, derivative slope calculation, squaring non-linearity, moving-window integration, and dual adaptive thresholds.
Measurable Outcome
Achieved sub-2ms per-frame pipeline latency with 99.2% R-peak sensitivity and accurate extraction of SDNN, RMSSD, and PR/QT intervals.