Söderman M, Picard C, Ericson K
Department of Neuroradiology, Karolinska Hospital, Stockholm, Sweden.
Neuroradiology. 1998 May;40(5):277-82. doi: 10.1007/s002340050585.
An algorithm for correction of the geometrical distortion in digital subtraction angiography (DSA) images was developed. Originally invented for 3D X-ray angiography, the algorithm was implemented in a computer program designed to fulfil the specific needs of stereotaxic DSA. The algorithm is based on transformation of an image of a grid from a distorted image back into its original pattern. The same transformation is then applied pixel-by-pixel to the angiographic images, which are acquired in direct conjunction with the grid image, without moving the gantry. The algorithm was tested in phantom studies and in the clinical situation with seven patients in ten examinations. Comparisons were made between co-ordinate determinations made on conventional full-size cut film and those performed on uncorrected and corrected DSA images, using 30- and 23-cm fields of view. With our method of measurement we could not shown any remaining geometric distortion in the corrected DSA images. This distortion correction can, if properly applied, be used for high-precision stereotaxic DSA.
开发了一种用于校正数字减影血管造影（DSA）图像中几何失真的算法。该算法最初是为三维X射线血管造影发明的，已在一个旨在满足立体定向DSA特定需求的计算机程序中实现。该算法基于将网格图像从失真图像转换回其原始图案。然后，将相同的变换逐像素应用于与网格图像直接同时采集的血管造影图像，而无需移动机架。该算法在模型研究以及对7例患者进行10次检查的临床情况下进行了测试。使用30厘米和23厘米视野，对在传统全尺寸切割胶片上进行的坐标测定与在未校正和校正后的DSA图像上进行的坐标测定进行了比较。通过我们的测量方法，我们未能在校正后的DSA图像中显示出任何残留的几何失真。如果正确应用，这种失真校正可用于高精度立体定向DSA。
