AUTOMATIC ANALYSIS AND VISUALIZATION OF MULTILEAD LONG-TERM ECG RECORDINGS

Long-term ECG data collected during physical exercises (stress tests) and ordinary daily activity (holter recordings) are necessary for extended diagnosis of ischemia and transient cardiac arrhythmias. Owing to the computer-assisted systems, the cardiologist decision is considerably facilitated by fast methods for automated ECG analysis. The main requirements to such sophisticated system are high accuracy for identification of each normal and abnormal cardiac contraction, correct measurements over the beats ' waveform, as well as adequate visualization of the results in the context of easy diagnostic interpretation. This work presents a PC-based application for automatic analysis of multilead ECG recordings. It implements fast algorithms for multilead QRS detection and classification of the QRS waveforms considering each user-selected channel. Categorized long-term diagrams are plotted showing the timing of each consecutive beat. The RR-interval trend is also provided. Signal-averaged technique is applied to obtain noise-free and stable P-QRS-T pattern at preset time-intervals during physical exercise. Continuos measurements of the P-QRS-T patterns allow the tracing of any changes provoked over time. Adequate graph representation facilitates the iterative visual identification of such changes. The basic software is opened for cardiologist-assistant supplements