“…From Table 1, it is observable that most authors developed deep CNN models [35], [37], [40], [41], [43], [46], [47], [57], [59], [61] for the automated classification of MI/CAD/CHF and normal classes while few authors developed hybrid deep models using CNN [39], [42], [45], [51], [53], [18]. Fewer authors employed other deep models such as the deep belief model [48], autoencoders [49], deep multilayer perceptron [52], deep ensemble models [56], deep neural network [60] and long-short term memory model(LSTM) [54] and conventional machine learning classifiers such as artificial neural networks [33], [34], [36], [39], [58] for the classification. High classification accuracies of about 95% were achieved when integral features were extracted using neural networks in [33] and from CNN models [35] [47].…”