"Spectral analysis and synthesis for non-stationary random signals"Torresani, BrunoIn the statistical signal processing or time series analysis literature, spectral analysis generally corresponds to the problem of estimating the power spectrum of a stationary signal from discrete, finite length realization. However, most real world signals are not stationary and such tools do not apply any more. This talk will address the non-stationary spectral analysis problem, with a focus on classes of random signals generated by deformation of stationary random processes. We will describe an appraoch, called JEFAS. In addition to spectral analysis, we will also address a dual synthesis-based approach, which models random signals as time-frequency synthesis from a time-frequency random field, with prescribed prior probability distribution. |
« back