Wavelet (Time-Scale) Analysis in Biomedical Signal Processing

Abstract

Signals recorded from the human body provide information pertaining to its organs. Their characteristic shape, or temporal and spectral properties, can be correlated with normal or pathological functions. In response to dynamical changes in the function of these organs, the signals may exhibit time-varying as well as nonstationary responses. Time-frequency and time-scale analysis techniques are well suited for such biological signals. Joint time-frequency signal analysis techniques include shortterm Fourier transform and Wigner-Ville distribution, and related reduced interference distribution. Joint time-scale analysis includes continuous and discrete, orthonormal and non-orthonormal wavelets. These techniques find applications in the analysis of transient and time-varying events in biological signals. Examples of applications include cardiac signals (for detection of ischemia and reperfusion-related changes in QRS complex, and late potentials in ECG) and neurological signals (evoked potentials and seizure spikes).