Toronto Metropolitan University
Browse

Despeckling Ultrasound Images with Decorrelated Compounding in the Spatial Frequency Domain

Download all (2.6 MB)
preprint
posted on 2024-11-27, 01:38 authored by Na ZhaoNa Zhao, Donald Kondi, Vanessa Hoang, Krista Ariello, Darren A. Yuen, Eno Hysi, Yuan XuYuan Xu

Speckle variance in ultrasound images limits the detection of low-contrast targets. Recently, a decorrelated compounding method in synthetic transmit aperture ultrasound imaging was proposed to reduce speckle variance. This paper extended the decorrelated compounding method beyond synthetic transmit aperture ultrasound imaging to any radio-frequency ultrasound images. A spatial-frequency-domain compounding method is proposed here by dividing the spectrum of ultrasound images into overlapped sub-domains to produce multiple sub-images for decorrelation and compounding. The image quality metrics, such as speckle signal-noise-ratio, contrast, contrast-noise-ratio, spatial resolutions, and lesion signal-noise-ratio,  were calculated to compare the proposed method and the delay-and-sum method.  The proposed method improved the contrast-noise ratio over the delay-and-sum method at the cost of spatial resolution loss. Overall, the proposed method improved lesion detectability in terms of lesion signal-noise-ratio by more than 57\% in numerical simulations and experimental studies.  The proportion correct  of detecting simulated low-contrast (-1 dB) lesions increased  from the 67\% in the the delay-and-sum method to 95\% in the proposed method in the two-alternative forced-choice studies on human observers. Structures in the {\it in vivo} clinical image were identified more easily in the proposed method than in the delay-and-sum method.

History

Language

English

Usage metrics

    Physics

    Licence

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC