Band Selective Volterra Filter for Nonlinear Ultrasound Imaging

Abstract

Nonlinear ultrasound imaging methods such as tissue harmonic imaging (THI) and pulse inversion (PI) have improved image quality, but have shortcomings that reduce their specificity. These limitations stem from loss of axial resolutions due to bandwidth restrictions, sensitivity to motion artifacts due to multiple transmissions per image line, loss of dynamic range due to echo subtraction etc. Several variants of these methods to overcome their limitations are being investigated by research groups worldwide. In this paper, we have developed a band selective quadratic Volterra filter capable of extracting nonlinear echo components throughout the entire bandwidth for improved spatial and contrast resolution. The performance of the proposed method is compared with the existing imaging techniques such as PI and truncated SVD (TSVD) quadratic kernel design method. Imaging results from a quality assurance phantom without ultrasound contrast agents (UCA) as well as an in vivo porcine kidney data with UCA are used to demonstrate applicability of the proposed Volterra filter as a nonlinear echo separation model irrespective of the source of nonlinearity. The new band selective kernel design method is shown to enhance contrast and lateral resolution while preserving axial resolution without the need for multiple transmission per A-line.