Deep learning for fast denoising filtering in ultrasound localization microscopy

Abstract

<i>Objective.</i>Addition of a denoising filter step in ultrasound localization microscopy (ULM) has been shown to effectively reduce the error localizations of microbubbles (MBs) and achieve resolution improvement for super-resolution ultrasound (SR-US) imaging. However, previous image-denoising methods (e.g. block-matching 3D, BM3D) requires long data processing times, making ULM only able to be processed offline. This work introduces a new way to reduce data processing time through deep learning.<i>Approach.</i>In this study, we propose deep learning (DL) denoising based on contrastive semi-supervised network (CS-Net). The neural network is mainly trained with simulated MBs data to extract MB signals from noise. And the performances of CS-Net denoising are evaluated in both<i>in vitro</i>flow phantom experiment and<i>in vivo</i>experiment of New Zealand rabbit tumor.<i>Main results.</i>For<i>in vitro</i>flow phantom experiment, the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of single microbubble image are 26.91 dB and 4.01 dB, repectively. For<i>in vivo</i>animal experiment , the SNR and CNR were 12.29 dB and 6.06 dB. In addition, single microvessel of 24<i>μ</i>m and two microvessels separated by 46<i>μ</i>m could be clearly displayed. Most importantly,, the CS-Net denoising speeds for<i>in vitro</i>and<i>in vivo</i>experiments were 0.041 s frame^-1and 0.062 s frame^-1, respectively.<i>Significance.</i>DL denoising based on CS-Net can improve the resolution of SR-US as well as reducing denoising time, thereby making further contributions to the clinical real-time imaging of ULM.