Self-organizing QRS-wave recognition in ECG using neural networks

Area of Science:

• Biomedical Engineering
• Artificial Intelligence
• Signal Processing

Background:

• Electrocardiograms (ECGs) are crucial for diagnosing cardiac conditions.
• Accurate QRS-wave detection is essential for ECG analysis.
• Existing QRS-wave recognition methods can be complex and computationally intensive.

Purpose of the Study:

• To develop a novel self-organizing neural network system for QRS-wave recognition in ECGs.
• To utilize the adaptive resonance theory (ART2) network for robust and adaptive pattern recognition.
• To achieve high precision in identifying Q and S points within the QRS complex.

Main Methods:

• A preprocessor identifies R points and segments ECGs into cardiac cycles.
• An ART2 network processes cardiac cycles to approximate Q and S point locations.
• A recognizer refines Q and S point detection within defined search regions.
• The ART2 network dynamically learns new QRS-wave patterns from incoming ECG data.

Main Results:

• The developed system effectively recognizes QRS-waves in ECG signals.
• The ART2 network demonstrates self-organizing capabilities, adapting to new ECG patterns.
• The system achieves an average recognition error of less than 1 millisecond for Q and S point localization.

Conclusions:

• The self-organizing ART2 network provides an accurate and adaptive solution for QRS-wave recognition.
• This system offers a promising approach for automated and precise ECG analysis.
• The low error rate highlights the potential clinical utility of this neural network system.