ABSTRACT The adaptive chirplet expansion (ACE) is proposed to characterize high-intensity, transient signals from circulating microemboli. The nonnegative adaptive spectrogram based on the ACE gives a compact representation of the microembolic signal (MES) in joint-time, frequency domain. The mean instantaneous power (MIP) and mean instantaneous frequency (MIF) of MES are estimated from the adaptive spectrogram. Then, several important characteristics of MES, such as embolus-to-blood ratio (EBR), half width maximum (HWM), and embolic signal onset (ESO), are computed from the MIP, and the frequency modulation is examined in the MIF. To validate the new method, we improved the simulation model of the audio Doppler ultrasound signal. Some MESs together with a Doppler ultrasound signal from carotid blood flow are simulated in the simulation study. As a comparison, the adaptive Gabor expansion (AGE) also is implemented on these simulated signals. The experimental results of the simulation study show that the new method, based on the ACE, outperforms the AGE-based method in MES characterization. The consistent conclusion has been confirmed by the clinical study on some clinical MESs.
© 2005, by The Institute of Electrical and Electronics Engineers, Inc. All rights reserved.