基于滤波反投影算法的血管内光声图像重建

Reconstruction of Intravascular Photoacoustic Images Based on Filtered Back-projection Algorithm

将目前在光声断层(PAT)成像中得到广泛应用的滤波反投影(FBP)重建算法应用到血管内光声(IVPA)成像中,提出一种简单快速的二维图像重建方法。首先,对组织产生的光声信号进行滤波、逆卷积和时域一阶求导的预处理;然后,针对IVPA在血管腔内封闭成像的特殊性,采用权重法将预处理后的光声信号数据对导管以外的成像区域沿弧线进行反投影,得到成像平面内每个网格点处的初始光声压。最后,得到反映血管壁组织结构形态的横截面灰阶图像。对仿真血管模型的实验表明,采用所提出的方法重建IVPA图像的结构,相似性指标(SSIM)可达到0.571 7。合理选择滤波函数、滤波截止频率以及测量位置数,可以提高IVPA重建图像的质量;对光声信号进行时域一阶求导处理,能有效地突出重建图像中的组织结构信息。该方法为后续图像重建算法的优化奠定基础。

A simple and fast image reconstruction method based on filtered back-projection （FBP） that has been widely applied in photoacoustic tomography （PAT） was proposed for intravascular photoacoustic （IVPA） imaging. First, the raw photoacoustic signals generated by the tissues xxere preprocessed by filtering, deconvolution and first-order derivation in time domain. Then, according to the specificity of IVPA scanning in the closed vascular lumen, the pre-processed photoacoustic signal data were back-projected along arcs with a weighting method to obtain the initial photoacoustic pressure of each grid point in the imaging region outside the catheter. Finally, grayscale images displaying the morphological structure of vascular cross-sections were constructed. The experimental results with a computer-simulated vessel phantom show that the structural similarity （SSIM） of the reconstructed image reached 0. 5717. The image quality was improved by setting a proper filter prototype, cut-off frequency and the number of measuring positions. The structural information of the vessel wall and plaque tissues can be effectively enhanced through first-order derivation performed to the raw photoacoustic signals. This method provides the foundation for subsequent optimization of image reconstruction algorithm.