Hayashi N, Sakai T, Kitagawa M, Inagaki R, Sadato N, Ishii Y, Nishimoto Y, Tanaka M, Fukushima T, Komuro H, Ogura H, Kobayashi H, Kubota T
Department of Radiology, Fukui Medical School, 23 Shimoaizuki, Matsuoka-cho, Yoshida-gun, Fukui 910-11, Japan.
Cardiovasc Intervent Radiol. 1998 Mar-Apr;21(2):138-41. doi: 10.1007/s002709900230.
Misregistration artifact is the major cause of image degradation in digital subtraction angiography (DSA). The purpose of this study was to evaluate the efficacy of a newly developed nonlinear geometric warping method to reduce misregistration artifact in DSA.
The processing of the images was carried out on a workstation with a fully automatic computerized program. After making differential images with a lapracian filter, 49 regions of interest (ROIs) were set in the image to be processed. Each ROI of the live image scanned the corresponding ROI of the mask image searching for the best position to match itself. Each pixel of the mask image was shifted individually following the data calculated from the shifts of the ROIs. Five radiologists compared the images produced by the conventional parallel shift technique and those processed with this new method in 16 series of cerebral DSA.
In 14 of 16 series (88%), more radiologists judged the images processed with the new method to be better in quality. Small arteries near the skull base and veins of low density were clearly visualized in the images processed by the new method.
This newly proposed method could be a simple and practical way to automatically reduce misregistration artifacts in DSA.
配准误差伪像是数字减影血管造影（DSA）中图像退化的主要原因。本研究的目的是评估一种新开发的非线性几何变形方法在减少DSA中配准误差伪像方面的效果。
图像的处理在配备全自动计算机程序的工作站上进行。使用拉普拉斯滤波器制作差分图像后，在待处理图像中设置49个感兴趣区域（ROI）。实时图像的每个ROI扫描蒙片图像的相应ROI，以寻找自身的最佳匹配位置。蒙片图像的每个像素根据从ROI的移位计算出的数据单独移位。五位放射科医生在16组脑DSA中比较了传统平行移位技术产生的图像和用这种新方法处理的图像。
在16组中的14组（88%）中，更多的放射科医生认为用新方法处理的图像质量更好。在新方法处理的图像中，颅底附近的小动脉和低密度静脉清晰可见。
这种新提出的方法可能是一种简单实用的自动减少DSA中配准误差伪像的方法。
