Classification of ECG Images Using Probabilistic Neural Network Based on Statistical Feature Analysis

Research on the analysis of ECG is mostly used for automating the diagnosis of different cardiac diseases. The ECG waveforms may differ for same patient to such extent that they are unlike each other and at same time alike for different types of beats. Many algorithms have been developed for the classification and detection of the ECG beat. In order to improve the accuracy of the ECG image feature extraction and classification system, the present research work proposes the use of different feature extraction methods. ECG image feature extraction and classification system, uses five feature extraction methods, namely wavelet decomposition, edge detection, gray level histogram, fast fourior transform and mean–variance, and probabilistic neural network classification. The objective of this present research work is to achieve the high accuracy and simplest classifiers related to extract the input features. An ECG image is classified by PNN using various feature extraction. The experimental results shows that wavelet decomposition gives a maximum accuracy compared to other feature extraction methods (Tayel MB, El-Bouridy ME (2008) ECG images classification using artificial neural network based on several feature extraction methods. International conference on computer engineering and system pp 113–115).