A comparison between image and signal sharpness-based axial localization of ultrasound scatterers

Abstract

Super-resolution ultrasound imaging deploys contrast microbubble (MB) tracking to delineate micro-vessels. The potential application spans to a large number of diseases which cause compromised vascular networks. Current super-resolution methods are mainly based on image processing. Sharpness-based localization is an alternative to such methods for scatterer localization in the axial direction, and can be implemented using both image and signal data. A 7-MHz, linear ultrasound transducer $(\lambda = 212 \mu \mathrm{m})$ and the Synthetic Aperture Real-time Ultrasound System (SARUS) were used to image a wire-target (point scatterer) at different depth positions. The method predicts a depth estimate and its difference from the true scatterer position demonstrates its accuracy. This average difference can be as low as $27.41 \mu \mathrm{m}($ or $\approx \lambda /8)$ for the image-derived sharpness and drops to $2.84 \mu \mathrm{m}($ or $\approx \quad \lambda /75)$ when the signals are used. These figures were calculated for a 8 mm depth range, which can be extended subject to further processing. The process of image formation involves interpolation and logarithmic compression that reduce the overall performance of the method when using image data. Such details may be significant when reconstructing micro-vessels of the order of tens of micrometres in diameter.