QRS Detection and Measurement Method of ECG Paper Based on Convolutional Neural Networks.

Area of Science:

• Biomedical Engineering
• Medical Imaging Analysis
• Artificial Intelligence in Healthcare

Background:

• Traditional electrocardiograph (ECG) analysis often involves manual interpretation or complex image-to-digital conversion processes.
• Accurate QRS complex detection is crucial for diagnosing cardiac arrhythmias and other heart conditions.
• Existing automated methods may struggle with direct analysis of ECG paper images.

Purpose of the Study:

• To develop and validate an end-to-end deep learning approach for QRS complex detection and measurement directly from ECG paper images.
• To leverage convolutional neural networks (CNNs) for automated analysis of visual ECG data.
• To compare the performance of the proposed method against conventional techniques.

Main Methods:

• An end-to-end system utilizing Faster-RCNN for direct QRS complex detection in ECG images.
• A subsequent CNN model for precise R-peak localization and measurement.
• Validation using clinical ECG data from the St.-Petersburg Institute of Cardiological Technics 12-lead Arrhythmia Database and real ECG paper from Peking University People's Hospital.

Main Results:

• Achieved a recall of 98.32% and a precision of 99.01% for QRS complex detection.
• Demonstrated a mean absolute error of 0.012 mV in R-peak measurement.
• Experimental results indicate superior performance compared to conventional ECG paper analysis methods.

Conclusions:

• The proposed CNN-based method offers a highly accurate and efficient solution for automated QRS complex detection and measurement directly from ECG paper.
• This approach eliminates the need for image-to-digital conversion, simplifying the analysis workflow.
• The findings suggest a promising direction for applying deep learning in the automated interpretation of visual ECG data.