基于压缩感知的超声逆散射成像研究  

Study of the ultrasound inverse scattering imaging based on compressed sensing

作　　者：花少炎[1] 丁明跃[1] 尉迟明[1] 

机构地区：[1]华中科技大学生命科学与技术学院,武汉430074

出　　处：《北京生物医学工程》2015年第1期24-31,37,共9页Beijing Biomedical Engineering

基　　金：国家重大科研仪器专项(2013YQ160551);十二五支撑计划(2012BAI13B020)资助

摘　　要：目的逆散射超声成像具有较高的空间分辨率和图像对比度,有着重大的临床应用价值。传统的基于逆问题理论的超声逆散射重建算法存在重建系统稳定性差,设备数据采集量大等问题。本文基于超声逆散射理论,利用压缩感知原理,对目标进行重建,以降低设备数据处理量,增强重建过程的稳定性。方法首先分别从时域和频率两个角度建立超声散射正向模型;再根据压缩感知原理,提出数据采集方案,获得投影观测数据;然后,利用目标的稀疏性,建立基于压缩感知的超声逆散射重建逆问题;最后,以囊肿体模和点目标为例,求解逆问题,重建图像。结果提出的基于压缩感知超声逆散射重建算法,对囊肿体模采样率降低50%,对点目标采样率降低76%。结论基于压缩感知的超声逆散射重建,能够降低设备数据处理量,与传统的延时叠加、合成孔径等成像算法相比,重建的图像具有更高的空间分辨率和对比度。Objective Ultrasound inverse scattering possesses high clinical application value due to its characteristics of high spatial resolution and contrast. Conventional reconstruction methods based on inverse theory usually deteriorate the system stability and need a large amount of sampled data. Based on inverse scattering theory,we reconstruct the object under compressed sensing （CS） framework to reduce the amount of data processing and enhance the stability of the reconstruction system. Methods Firstly, we formulate the forward scattering model from time and frequency domain respectively. Secondly, we propose the data acquisition scheme to obtain projection data. Thirdly,the inverse problem under CS theory is formulated based on the sparsity of the object. Finally, the inverse problem is solved, and the cyst and point objects are constructed. Results The CS approach can faithfully reconstruct the object, the sample data is reduced 50% and 76% for cyst and point object,respectively. Conclusions The CS approach can reduce the amount of data processing, enhance the stability of the inverse problem. Compared to conventional methods such as delay and sum,synthetic aperture,the CS approach improves spatial resolution and contrast.

关 键 词：压缩感知 逆散射 随机采样 稀疏性 

分 类 号：R318.04[医药卫生—生物医学工程] TP391.9[医药卫生—基础医学]