A new automated heart defect detection based on signals and systems concept

In this paper, a new automated early-stage detection of heart defect is introduced. Based on the proposed technique, human heart is modeled as a pump system where ECG signal and PCG signal are defined to be command (input) and response (output) of the system, respectively. The ratio of discrete Fourier transform of both signals is employed to determine the transfer function H_p(k) of a patientpsilas heart system. The early-stage heart defect detection is fundamentally based on the correlation between the standard transfer function H_s(k) and the patient heart transfer function H_p(k). The obtained correlation measurement is fed to be the input of neural network to make decision whether or not patientpsilas heart is normal. The proposed method provides satisfying results where negative and positive false rates approximately are 12.50% and 11.11%, respectively.