报告人：Lawrence H. Le
内容简介：Ultrasonic signals acquired by a transducer array are easily distorted by cross-coupling artefacts or crosstalk due to inter-communication among individual elements of the array. The crosstalk are coherent signals, which interfere with the real signals traveling through the medium and their removal is essential, prior to any data analysis, to provide an accurate interpretation of the signals. In this study, adaptive noise cancellator (ANC) was used to estimate the crosstalk-free signals. The filter used the acquired signals and reference crosstalk as the input to adaptively filter the reference crosstalk and subtract them from the input signals to provide the estimate. Instead of operating the filter in the time-offset (t-x) domain, the t-x data was first Radon-transformed to the time interceptslowness (tau-p) domain and ANC was used to eliminate crosstalk in the transformed domain. Simulated data was used to validate the accuracy and efficacy of the modified ANC. The filter was also applied to phantom and bone-plate data and the filtered signals were compared with those of the classical ANC.