Time-domain diffuse optical tomography (TD-DOT) uses near-infrared pulsed lasers as light sources to measure time-varying exitance on the boundary of the target. These are used to estimate optical properties of the imaged target. Several integral-transform-based moments of the time-resolved data have been utilized in TD-DOT, the most common being the mean time of flight and variance. Recently, it has been shown that Fourier transforming the time-domain data to frequency domain enables utilization of these data at one or several frequencies, producing equally as good estimates as the whole time-domain data. In this work, we present a systematic comparison of the usage of the temporal moments and Fourier transformed data in TD-DOT. Both absolute and difference imaging are evaluated using numerical simulations. The simulations show that utilizing temporal moments and Fourier transformed data in TD-DOT provides good quality reconstructions with a good estimation accuracy. These estimates are improved if more than one data type is used. Furthermore, the simulations show that the frequency-domain computations enable computationally cheaper and straightforward implementation of the inverse solver when compared to the temporal moments.

中文翻译：

---

在时域漫射光学层析成像中评估时间矩和傅立叶变换数据

时域漫射光学层析成像（TD-DOT）使用近红外脉冲激光作为光源来测量目标边界上随时间变化的出射量。这些用于估计成像目标的光学特性。TD-DOT中已经利用了几个基于积分变换的时间分辨数据矩，最常见的是平均飞行时间和方差。近来，已经表明傅立叶将时域数据变换到频域使得能够在一个或几个频率上利用这些数据，从而产生与整个时域数据一样好的估计。在这项工作中，我们对TD-DOT中的瞬时矩和傅立叶变换数据的用法进行了系统的比较。绝对成像和差值成像均使用数值模拟进行评估。仿真表明，在TD-DOT中利用时间矩和傅立叶变换数据可提供具有良好估计精度的高质量重建。如果使用一种以上的数据类型，这些估计将得到改善。此外，仿真表明，与时间矩相比，频域计算使逆求解器的计算更便宜，更直接。