Ultrasonography is non-invasive and painless. In Ultrasonography the images are often affected with Speckle noise. It is a multiplicative noise. To help the doctors to identify the abnormalities properly there are several methods to diagnose as speckle is a major problem. This paper gives details about popular spatial domain, transform domain, CNN techniques for despeckling in ultrasound images. Transform domain methods like Wavelet methods, Curvelet methods, Bayes Shrink methods are prominent among many researches. Deep learning based methods are evolving like DnCNN, ECNDNet etc. for efficient despeckling. An overview of the methods is given here with certain measurement parameters like PSNR, MSE.

超声检查是非侵入性和无痛的。在超声检查中，图像经常受到斑点噪声的影响。它是一种乘法噪声。为了帮助医生正确识别异常，有几种方法可以诊断斑点是一个主要问题。本文详细介绍了流行的空间域、变换域、CNN 技术在超声图像中的去斑。小波方法、曲线方法、贝叶斯收缩方法等变换域方法在许多研究中都很突出。基于深度学习的方法正在发展，如 DnCNN、ECNDNet 等，以实现高效的去斑。这里给出了这些方法的概述，其中包含某些测量参数，如 PSNR、MSE。