In quantitative phase imaging, sensitivity is a key measure of system reproducibility. Despite continuous experimental breakthroughs in achieving highly sensitive detection, in-depth studies of theoretical constraints on sensitivity are inadequate and comparisons between different techniques are difficult. In this paper, we introduce the method to evaluate the sensitivity of phase shifting interferometry which is a major category of quantitative phase imaging techniques. The method discusses in detail several key concepts of sensitivity evaluation, including a general three-level evaluation framework, a complete interference signal model, Cramér-Rao bound and algorithm sensitivity, algorithm and system efficiencies, as well as energy efficiency of an algorithm. In discussions of specific phase shifting algorithms, we focus on the shot noise-limited model. This simplified model not only reflects the rapid developments in modern detectors that are often dominated by shot noise, but also permits the calculation of theoretical sensitivities based on measured data, which is important in evaluating experimental performance. As examples, we study several common phase shifting interferometric techniques. The results of different techniques are compared to provide insights into algorithm optimization and energy efficiency of sensitivity. A normalized algorithm sensitivity table is also provided for readers to conveniently estimate their system's algorithm sensitivity and compare against experiments.

中文翻译：

---

相位灵敏度评估及其在相移干涉测量中的应用

在定量相位成像中，灵敏度是衡量系统重现性的关键指标。尽管在实现高灵敏度检测方面不断取得实验突破，但对灵敏度的理论约束的深入研究还不够，不同技术之间的比较也很困难。在本文中，我们介绍了评估相移干涉测量灵敏度的方法，相移干涉测量是定量相位成像技术的主要类别。该方法详细讨论了灵敏度评价的几个关键概念，包括通用的三级评价框架、完整的干扰信号模型、Cramér-Rao界和算法灵敏度、算法和系统效率以及算法的能量效率。在讨论具体的相移算法时，我们专注于散粒噪声限制模型。这种简化的模型不仅反映了通常由散粒噪声主导的现代探测器的快速发展，而且还允许根据测量数据计算理论灵敏度，这对于评估实验性能很重要。作为例子，我们研究了几种常见的相移干涉技术。比较不同技术的结果，以深入了解算法优化和灵敏度的能量效率。还提供了归一化的算法灵敏度表，供读者方便地估计其系统的算法灵敏度并与实验进行比较。但也允许根据测量数据计算理论灵敏度，这对于评估实验性能很重要。作为例子，我们研究了几种常见的相移干涉技术。比较不同技术的结果，以深入了解算法优化和灵敏度的能量效率。还提供了归一化的算法灵敏度表，供读者方便地估计其系统的算法灵敏度并与实验进行比较。但也允许根据测量数据计算理论灵敏度，这对于评估实验性能很重要。作为例子，我们研究了几种常见的相移干涉技术。比较不同技术的结果，以深入了解算法优化和灵敏度的能量效率。还提供了归一化的算法灵敏度表，供读者方便地估计其系统的算法灵敏度并与实验进行比较。