International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 109 - Number 2 |
Year of Publication: 2015 |
Authors: P. S. Gawande, S. A. Ladhake |
10.5120/19158-0601 |
P. S. Gawande, S. A. Ladhake . Artificial Neural Network based Electrocardiogram Classification for Biometric Authentication. International Journal of Computer Applications. 109, 2 ( January 2015), 6-9. DOI=10.5120/19158-0601
In this study, Multilayer Perceptron (MLP) neural network model is used for the classification of electrocardiogram (ECG) signals. Out of twelve leads recorded lead II, single lead, is used for analysis. ECG samples from eight normal individuals are recorded regularly almost every month for thirty six months. Input to the network is feature vector matrix of ten features and seven statistical and three morphological features are extracted. MLP networks with single hidden layer are trained with three runs and one thousand epochs. The testing results demonstrate that the neural network is effective tool for this application and the accuracy of 99.76 is observed during experimentation.