Windowed Nonlinear Energy Operator-based First-arrival Pulse Detection for Ultrasound Transmission Computed Tomography

A windowed nonlinear energy operator (wNEO)-based algorithm is proposed to detect the first-arrival pulse of ultrasound transmission computed tomography (UTCT). In UTCT, the detection of first-arrival pulse is critical to derive imaging indices such as attenuation coefficients and time-of-flight (TOF). Currently, various algorithms utilizing a threshold, maximum, wavelet, etc. have been tested for the detection of the pulse, but each technique suffers from its own intrinsic limitations, failing in detecting the first-arrival pulse reliably. Especially the most widely used wavelet technique, which is also considered as the most effective technique to handle frequency-dependent attenuation, dispersion, and noise of ultrasound, often fails to detect the first-arrival pulse due to the inconsistency of ultrasound pulse shapes with respect to a best mother wavelet. Furthermore, the wavelet technique is impractical for real-time processing. In this work, we propose the use of wNEO for the first time to detect the first-arrival pulse and show its outperforming results in comparison to those conventional techniques through simulation and experimental data with a pair of 5MHz ultrasound transmit-receive transducers.