基于水平集的光学层析图像重建

A Reconstruction Method for Optical Tomography Based on Level Set Technique

为解决重建图像中异常区域边界不规则这一问题,提出一种基于水平集的光学层析图像重建方法。该方法在背景区域光学参数已知情况下,可以使重建异常目标区域的位置更加接近其真实位置,且实现了异常区域位置和光学参数的同时重建。采用扩散方程作为光子传输模型,基于有限元方法实现正向模型的数值求解,用零水平集曲线表示异常区域边界。实验结果表明,该方法不仅能够更好的实现异常目标区域的定位而且能够方便的追踪目标区域拓扑结构的变化,与传统重建方法相比,极大提高了重建图像质量和重建速度。

An optical tomography image reconstruction method based on level set technique is proposed to solve the problem of irregular borders in the abnormal area of the image reconstruc- tion. Optical parameter of background known, the target area shown by the result of reconstruc- tion is closer to actual value, and abnormal boundaries and optical properties can be recovered si- muhaneously. In this paper, diffusion equation is adopted as the forward model, and its numeri- cal solution is obtained by the finite element method. In the reconstruction process, zero level set is used to represent boundaries of abnormal regions. Experiment results show that this method can not only acquire a better position for target area but also make it easy to follow the topology change for target area. Compared with the traditional reconstruction methods, the image quality and reconstruction speed can be improved greatly.