基于纹理分析的自适应超声影像衰减补偿

针对人体组织衰减超声波所引起的超声图像整体亮度不均匀、组织结构模糊、难以辨别问题,首次提出一种基于图像纹理分析的自适应超声衰减能量补偿技术.首先描述超声波在软组织传播时能量的衰减模型,并引入组织的声衰减系数以及深度方向衰减能量的补偿量等概念.在介绍2维共生矩阵以及1维和差直方图表示的图像纹理参数后,阐述了准确估计组织声衰减系数及合理补偿超声衰减能量的基本原理.最后,通过已知声衰减系数的仿组织超声体模和人体组织超声影像做试验验证.实验结果不仅估计出准确的组织声衰减系数,亦可得到灰度值一致的超声影像.

Adaptive Ultrasound Image Attenuation Compensation Based on Texture Analysis

In order to get uniform ultrasound images with clear structure, a novel and adaptive ultrasound attenuation compensation method based on image texture analysis was investigated. The attenuation process of acoustic signals transmitted in soft tissue was modeled, with the introduction of tissue attenuation coefficient and depth dependent attenuation compensation. After given the 2-D co-occurrence matrices and 1-D sum and difference histograms based image texture parameters, the basic idea to precisely estimate tissue attenuation coefficient and properly compensate the attenuated ultrasound power was proposed. Algorithms were verified both in tissue-mimicking phantom with known attenuation coefficient and in vivo ultrasound images acquired from a clinical ultrasound system. Testing results presented accurate tissue attenuation coefficients and also uniform attenuation compensated ultrasound images.