Estimation for Stochastic Inverse Problems

当期望信号或图像的模糊版本影响模拟测量数据的随机过程的统计量时，就会出现随机逆问题。例子包括体内放射性示踪剂的量子限制成像，天文学中的伽马射线源成像，以及在弱光摄影中形成图像。约翰·霍普金斯大学的艾伦·卡尔（Alan Karr）正在进行一个联合小组研究项目，以研究解决这些问题的新算法。识别映射到高度并行计算体系结构的算法是研究的一个重要目标。

Stochastic inverse problems arise when a blurred version of a desired signal or image influences the statistics of a random process that models measured data. Examples include the quantum- limited imaging of radioactive tracers in the body, imaging gamma-ray sources in astronomy, and forming images in low-light photography. A joint group research program is being pursued with Alan Karr at the Johns Hopkins University to investigate new algorithms for these problems. Identifying algorithms which map onto highly parallel computational architectures is an important objective of the research.