Douiri Abdel, Schweiger Martin, Riley Jason, Arridge Simon
Department of Computer Science, University College London, Gower Street, London WC1E 6BT, UK.
Opt Lett. 2005 Sep 15;30(18):2439-41. doi: 10.1364/ol.30.002439.
We formulate a solution to the diffuse optical tomography (DOT) inverse problem as the minimization of an energy functional of the solution and the data. For the solution prior we introduce a local diffusion regularization potential with a threshold based on robust statistics (the Hubert function). We compare results on simulated data for the Hubert function and two other standard regularization functionals, Tikhonov and total variation.
我们将扩散光学断层扫描（DOT）逆问题的解决方案表述为求解能量泛函与数据的最小化。对于解的先验信息，我们基于稳健统计（休伯函数）引入了一个带阈值的局部扩散正则化势。我们比较了休伯函数以及另外两种标准正则化泛函（蒂霍诺夫泛函和全变差泛函）在模拟数据上的结果。
