| Popis: |
In this research, an efficient technique has been developed to classify the five abnormal beat (Afonso et al., IEEE Trans Biomed Eng 46:192–202 (1999) [1]; Kohler et al., IEEE Eng Med Biol Mag 21:42–57 (2002) [2] signals which includes Left Bundle Branch Block beat (LBBB), Right Bundle Branch Block beat (RBBB), Premature Ventricular Contraction (PVC), Atrial Premature Beat (APB), and Nodal (junction) Premature Beat (NPB) along with the normal beat. The proposed technique is composed into three stages, (1) preprocessing (2) Hybrid feature extraction (3) classifier. For efficient feature extraction, hybrid feature extractor is used. Hybrid feature extraction is done in two steps, (i) Morphological-based feature extraction (ii) Haar wavelet-based feature extraction. Once the features are extracted, a Feed Forward Neural Network (FFNN) classifier classifies the beat signal. Artificial Bee Colony (ABC) combined with genetic algorithm has been used for training the neural network. A best crossover rate has been chosen in order to achieve higher accuracy. The proposed technique gives an accuracy of 81 %, sensitivity of 75 %, and specificity of 79 % for the crossover rate of 0.8. |
