摘要:
An image reconstruction method for reconstructing an image
ƒ min representing a region of investigation within an object (1), comprises the steps of providing detector data ( y i ) comprising Poisson random values from an X-ray transmission measurement using an X-ray source (10) and a detector device (20), said detector data ( y i ) being measured at an i -th of a plurality of different pixel positions of the detector device (20) relative to the object (1), and reconstructing the image
ƒ min based on the detector data ( y i ), said reconstructing step including a procedure of minimizing a functional (I) wherein ƒ is a current test image used for minimizing the functional F( ƒ ), (II) is a maximum-likelihood risk functional for Poisson statistics, said parameters μ i being transmission projections of the test image ƒ , said projections being computed according to Beer-Lambert's law at the i -th pixel position relative to the X-ray source (10), | T -1 ƒ | p is a sparsity enforcing functional including the l p norm of vector T -1
ƒ with 0 ≤ p -1 ƒ being a sparse or compressive representation of ƒ in a (bi-) orthogonal basis T , and a is a calibration factor, wherein the image ƒ
min represents the global minimum of the functional F ( ƒ ).
摘要翻译:用于重构表示物体内的调查区域的图像fmin的图像重建方法包括:使用X射线源和检测器装置的X射线透射测量提供具有泊松随机值的检测器数据(yi),检测器数据 yi)在检测器装置的多个不同像素位置的第i个相对于对象被测量,并且基于检测器数据(yi)重建图像fmin,重建步骤包括使功能F (f _)= 1kΣi = 1 k笨(μi - yi logüμμ)+ aT - 1 f _p其中f是用于最小化的当前测试图像 功能F(f)。 图像fmin表示F(f)的全局最小值。