Back-projection method with fast time-delay correction for photoacoustic tomography reconstruction based on a focused sound field model

Xiao-Fei Luo,Bo Wang,Jia-Ying Xiao,Kuan Peng

doi:10.7498/aps.71.20212019

Abstract

Photoacoustic tomography is a rapidly developing imaging technology, which can provide structural and functional information of biological tissues. It can integrate the advantage of high optical contrast from optical imaging with those from high penetration depth from ultrasound imaging. However, the existing back-projection algorithm in photoacoustic tomography equates the ultrasonic transducer scanning around the target to a point detector, which leads to a notable tangential blur in the eccentric imaging regions, and thus seriously degrading the image quality. In this paper, we propose a novel photoacoustic tomography reconstruction algorithm, which employs a focused sound field equivalent model to overcome the sound field distortion caused by the transducer’ finite aperture effect and can quickly and effectively restore the elongated tangential resolution in the eccentric imaging regions. Simulation results show that for the target with a diameter of 5 mm and a distance of 6 mm from the rotation center, the tangential resolution is improved by at least twice. Experimental results show that this method can effectively restore the image tangential blur in the off-center regions, where the tiny structures of complex targets can be detected. This new method provides a valuable alternative to the conventional back-projection method and plays an important guiding role in the design of photoacoustic tomography systems based on circle/sphere scanning.

Highlights

Simulation results with different reconstruction methods when the transducer central frequency varies. (a)-(d) The results with the conventional point-like method when the central frequency are 1 MHz, 3 MHz, 5 MHz and 10 MHz, respectively. (e)-(h) The corresponding results with the model of the infinite element size. (i)-(l) The corresponding results with the model of focused sound field
The simulated lateral profiles for the four targets obtained from 3 MHz central frequency ultrasonic transducer
The locations of the four points are 0 mm, 2 mm, 4 mm, and 6 mm, respectively

Summary

Introduction

在 PAT 中,通常对整个样本使用宽场照明,假设每个换能器覆盖整个目标图像域,并通过围绕样品的单换能器扫描或使用换能器阵列来采集超声信号.然后利用图像重建方法一次性恢复整个区域的光吸收分布[10]. PAT 扫描的几何形状可以是线性的、圆形的或球形的,也可以是上述几何形状的组合.与线性扫描[11]相比,圆形(二维成像)[9]和球形(三维成像)[10]扫描可以覆盖更完整的目标视角,以提供更好的成像质量.除了优化扫描几何结构外,快速准确的图像重建算法也是 PAT 研究的热点[12]. 目前,已经有几种方法来改善这种被拉伸的切向分辨率.很多研究人员提出了在基于圆周扫描的 PAT 中使用聚焦换能器并结合虚拟点探测器的方法[18]. 但据报道这些方法不能可靠地重建目标的高频部分[19].为了克服“有限孔径效应”, 一些研究者采用了基于各种最优化模型的图像重建方法,结合换能器的几何形状和频率响应对图像进行重建,但其计算成本太高.反卷积作为一种图像后处理方法也有可能提高切向分辨率,但作为一种典型的逆方法,它会引起较强的图像噪声[20, 21].因此,仍然需要发展新的 PAT 重建算法来克服基于圆形/球面扫描的 PAT 中有限换能器尺寸的影响.

Results

Conclusion