Blokland J A, Vossepoel A M, Bakker A R, Pauwels E K
Department of Diagnostic Radiology, University Hospital, Leiden, The Netherlands.
Eur J Nucl Med. 1989;15(2):87-92. doi: 10.1007/BF00702625.
Several algorithms for quantitative analysis of 201Tl scintigrams require operator assigned elliptical regions. Therefore these algorithms are hampered by inter- and intra-observer errors. In this paper we present a robust algorithm to determine automatically contours with elliptical shapes. The algorithm is based on the Hough transformation and can determine the parameters describing an ellipse even in low signal-to-noise ratio images and when the contour is only partially visualized such as in non-perfused regions of the left ventricular myocardium. As long as the number of parameters to describe the shape of the contour is not too large, the same kind of algorithm can also be used for differently shaped contours, and for 3D elliptical contours as in SPECT.
几种用于铊-201闪烁图定量分析的算法需要操作人员指定椭圆形区域。因此,这些算法受到观察者间和观察者内误差的影响。在本文中,我们提出了一种强大的算法,可自动确定椭圆形轮廓。该算法基于霍夫变换,即使在低信噪比图像中,以及当轮廓仅部分可见时,如在左心室心肌的非灌注区域,也能确定描述椭圆的参数。只要描述轮廓形状的参数数量不是太大,同样类型的算法也可用于不同形状的轮廓,以及用于单光子发射计算机断层扫描(SPECT)中的三维椭圆形轮廓。