International Research Journal of Education and Technology

Automatic electrocardiogram (ECG) signal classification is essential for the diagnosis of dangerous heart conditions. Here, first the noise is removed from the digitized ECG signal. After that QRS complex, T and P waves are detected and also delineated using different amplitude and threshold values. Finally the required features are extracted and then the ECG beats are classified into different heart abnormalities. In this paper, Myocardial Infarction, Premature Ventricular Contraction, Ventricular Tachycardia, Supra Ventricular arrhythmias, ST deviation, Ventricular Fibrillation are classified using Feed Forward Artificial Neural Network Classifier. Conventional segmentation methods present some limitations such as, need of definition of large number of parameters and lack of obvious way to tune all these parameters. Optimization is used to adjust the parameters of the delineator in order to minimize the cost function and it is able to find global solution in high dimensional search space. Particle Swarm Optimization (PSO) and Artificial Bee Colony Optimization (ABC) are used here.The databases are extracted from MIT-BIH, EURO VFDB, SVDB, MIT ST Change, CUVTDB databases. The dataset is first trained, validated and tested after which it is analyzed by using the Pattern Recognition Toolbox in MATLAB. The Artificial Neural Network achieved 81.1%% accuracy in its training phase before optimization and 90.5% with ABC and 92.9% with PSO. 

Key Words:  ECG, PSO, ABC, and Wavelet Transform