Signature of complexity in time–frequency domain

Abstract

We propose a time–frequency based complexity to measure disorder in the long term dynamics of a signal. The disorder is characterized by defining an wavelet spectrogram space in a multiscale coordinate system. The multiscale coordinate system is formed with Wavelet coefficient of the signal. Further, an Weighted entropy measure is implemented to quantify the aforesaid disorder. Numerical results support the proposed method. The proposed entropy is successfully applied to distinguish the ECG signals of normal healthy person and congestive heart failure patients.