摘要
斑点噪声是超声成像机制引起的固有噪声形态,它给超声图像的特征提取、识别和分析造成了困难.基于各向异性扩散方程的滤波技术能够有效地抑制斑点噪声,提出了一种基于各向异性扩散方程的医学超声图像降噪方法,该方法改进了林石算子,设置了扩散门限,并将各向异性扩散方程的常规扩散方向扩展为8方向.仿真医学超声图像降噪实验表明了该方法的滤波和保边性能都明显优于经典的PM方程和林石算子,而该方法在实际医学超声图像降噪上的成功应用表明了其临床应用价值.
Speckle noise is one of the inherent noise patterns evoked by the mechanism of ultrasound imaging, which brings great difficulties to feature extraction, recognition and analysis. Techniques based on the anisotropic diffusion equation can effectively suppress the speckle noise. An algorithm using the anisotropic diffusion equation was proposed to suppress the speckle noise of a medical ultrasound image. This method improves LinShi operator and enhances the conventional method by setting the threshold value and including diagonal edges. The simulation results showed that this method is superior to the classical PM equation and LinShi operator at image smoothing and edge preserving, and de-noising for real ultrasound image indicates that this method has high potential for chnical apphcations.
出处
《山东大学学报（工学版）》
CAS
北大核心
2009年第5期38-42,共5页
Journal of Shandong University（Engineering Science）
基金
国家自然科学基金资助项目(60571040)
山东省优秀中青年科学家奖励基金资助项目(2005BS1006)
关键词
医学超声图像
斑点噪声
各向异性扩散
medical ultrasound image
speckle noise
anisotmpic diffusion
