Quantitative resolution of ultrasound images

We discuss here the use of spectral estimation algorithms for biomedical ultrasound imaging. B-mode ultrasound imagery is based on computing the envelope of acoustic individual radio-frequency (RF) lines. An envelope estimator is thus needed to improve spatial resolution and reduce speckle. There is an analogy between spectral analysis and envelope estimation. Indeed, the signal envelope can be seen as the power spatial density. This study evaluated the potential of high-resolution (HR) spectral estimation techniques in the resolution enhancement of ultrasound images. Several spectral estimation methods were therefore compared using the Monte-Carlo method on a synthetic image for different noise and standard deviation conditions. The paper also illustrates their performance on 20 MHz in-vivo ultrasound imaging of a skin lesion.