Publication

Abstract : Electrocardiogram (ECG) is the simplest test, used for checking the heart’s rhythm and measuring the heart’s electrical activity. This method is the most probably used technique for identifying heart diseases because of its non-invasive nature. The patient’s status is identified by handling the irregularities in the ECG signals. For computation in the existing system, they use signal processing methods which result in time consumption in real time. In existing system, it is hard to predict the detection of heartbeat early and accurately. In this work, the heartbeat is classified using ECG signal by proposed Chronological Golden Search Optimisation (CGSO)-based Deep Learning (DL). Here, the CGSO algorithm is developed by hybridising the Chronological concept with the Golden search optimisation (GSO) algorithm, which is utilised to train SqueezeNet. After the acquisition of ECG signals, pre-processing is done by a median filter. Then, the features are extracted, and this tends to the data augmentation process. Finally, the heartbeat is classified using SqueezeNet-enabled CGSO. The performance is analysed using ECG heartbeat categorisation data set. The proposed model obtained a specificity of 93.5%, sensitivity of 93.1% and precision of 89.8%. From the results, it is known that the proposed system offers more accurate results.