This paper aims to enhance the image quality in ultrasound images. The significant difficulties in the ultrasound image are the presence of speckle noise. Speckle is the granular noise, and this kind of noise produces a lot of challenges during medical diagnosis. Also, these kinds of problems degrade image quality. The proposed work overcomes this kind of issue with nonsubsampled contourlet domain‐based cross‐guided bilateral filtering. Initially, additive speckle noise is introduced with the log‐transformed Rayleigh distribution. This additive speckle noise component is to be decomposed through the non‐subsampled contourlet domain transform. Then the decomposed image is despeckled by the cross‐guided bilateral filter, and the cost function of this filter is minimized through adaptive galactic swarm optimization. The proposed filtering technique despeckles the additive noise in the noisy image. Finally, the decomposed image is reconstructed using the inverse transform of the nonsubsampled decomposition. The implementation of the proposed methodology is carried out in MATLAB simulation platform. Also, the performance of the proposed work is measured by few metrics such as structural similarity index matrix, mean structural similarity index matrix, mean square error, peak signal to noise ratio, Bhattacharyya coefficient, and speckle index. The result of cost function optimization provides less distortion when compared to other techniques.

中文翻译：

---

具有交叉引导双边滤波器的非下采样Contourlet变换用于医学超声图像的去斑

本文旨在提高超声图像的图像质量。超声图像中的主要困难是斑点噪声的存在。斑点是颗粒噪声，这种噪声在医学诊断过程中会带来很多挑战。同样，这些类型的问题也会降低图像质量。拟议的工作通过基于非下采样contourlet域的交叉导引双边过滤技术克服了此类问题。最初，通过对数变换的瑞利分布引入加性散斑噪声。该附加的斑点噪声分量将通过非二次采样的contourlet域变换来分解。然后通过交叉引导的双边滤波器对分解后的图像进行散斑，并通过自适应银河群优化使该滤波器的代价函数最小化。所提出的滤波技术使噪声图像中的附加噪声去斑点。最后，使用非下采样分解的逆变换来重建分解图像。所提出方法的实现是在MATLAB仿真平台上进行的。而且，所提出的工作的性能是通过一些指标来衡量的，例如结构相似性指标矩阵，平均结构相似性指标矩阵，均方误差，峰信噪比，Bhattacharyya系数和散斑指标。与其他技术相比，成本函数优化的结果可提供较小的失真。所提出方法的实现是在MATLAB仿真平台上进行的。而且，所提出的工作的性能是通过一些指标来衡量的，例如结构相似性指标矩阵，平均结构相似性指标矩阵，均方误差，峰信噪比，Bhattacharyya系数和散斑指标。与其他技术相比，成本函数优化的结果可提供较小的失真。所提出方法的实现是在MATLAB仿真平台上进行的。同样，通过少量指标（如结构相似性指标矩阵，平均结构相似性指标矩阵，均方误差，峰信噪比，Bhattacharyya系数和斑点指标）来衡量所提出工作的性能。与其他技术相比，成本函数优化的结果可提供较小的失真。