ECG QRS Detection Methods

This repository contains MATLAB implementations of three different methods for QRS complex detection in ECG signals. The project uses the MIT-BIH Arrhythmia Database for testing and evaluation.

Methods Implemented

1. Zero-Crossing Method
2. Empirical Mode Decomposition (EMD) Method
3. Morphological Operations Method

Requirements

• MATLAB (version R2019b or later recommended)
• WFDB Toolbox for MATLAB
• MIT-BIH Arrhythmia Database

File Structure

Main Scripts

• main_crossing.m: Main script for QRS detection using the zero-crossing method
• main_emd.m: Main script for QRS detection using the EMD method
• main_morphological.m: Main script for QRS detection using morphological operations

Detection Functions

• detect_qrs_crossing.m: Implements the zero-crossing method for QRS detection
• detect_qrs_emd.m: Implements the EMD method for QRS detection
• detect_qrs_morphological.m: Implements the morphological operations method for QRS detection

Utility Functions

• evaluate_detection.m: Function to evaluate the performance of QRS detection algorithms

Setup

1. Clone this repository to your local machine.
2. Download the MIT-BIH Arrhythmia Database and place it in a folder named mit-bih-arrhythmia-database-1.0.0 in the project directory.
3. Install the WFDB Toolbox for MATLAB and ensure it's in your MATLAB path.

Usage

1. Open MATLAB and navigate to the project directory.
2. Run one of the main scripts:
   • main_crossing.m
   • main_emd.m
   • main_morphological.m
3. The script will process the ECG signal, detect QRS complexes, and display the results along with performance metrics (if annotations are available).

Customization

• To analyze a different record from the MIT-BIH Arrhythmia Database, change the record_name variable in the main script you're using.

Performance Metrics

The scripts calculate and display two performance metrics when annotations are available:

1. Sensitivity (Se): The proportion of actual QRS complexes that were correctly detected.
2. Positive Predictivity (+P): The proportion of detected QRS complexes that were actual QRS complexes.

These metrics are calculated using the evaluate_detection.m function.

Contributing

Contributions to improve the algorithms or add new QRS detection methods are welcome. Please feel free to submit pull requests or open issues for any bugs or suggestions.

License

This project is open-source and available under the MIT License.

Acknowledgments

• This project uses the MIT-BIH Arrhythmia Database from PhysioNet.
• The WFDB Toolbox for MATLAB is used for reading ECG signals and annotations.