Ostgaard S E, Gottlieb L, Toksvig-Larsen S, Lebech A, Talbot A, Lund B
Department of Orthopedics, National University Hospital, Copenhagen, Denmark.
J Biomech. 1997 Sep;30(9):993-5. doi: 10.1016/s0021-9290(97)00060-2.
The two-dimensional position of markers in radiographs for Roentgen Stereophotogrammetric Analysis (RSA) is usually determined using a measuring table. The purpose of this study was to evaluate the reproducibility and the accuracy of a new RSA system using digitized radiographs and image-processing algorithms to determine the marker position in the radiographs. Four double-RSA examinations of a phantom and 18 RSA examinations from six patients included in different RSA-studies of knee prostheses were used to test the reproducibility and the accuracy of the system. The radiographs were scanned at 600 dpi resolution and 256 gray levels. The center of each of the tantalum-markers in the radiographs was calculated by the computer program from the contour of the marker with the use of an edge-detection software algorithm after the marker was identified on a PC monitor. The study showed that computer-based image analysis can be used in RSA-examinations. The advantages of using image-processing software in RSA are that the marker positions are determined in an objective manner, and that there is no need for a systematic manual identification of all the markers on the radiograph before the actual measurement.
用于X射线立体摄影测量分析（RSA）的X线片中标记物的二维位置通常使用测量台来确定。本研究的目的是评估一种新的RSA系统的可重复性和准确性，该系统使用数字化X线片和图像处理算法来确定X线片中标记物的位置。对一个模型进行的四次双RSA检查以及来自六名患者的18次RSA检查（这些患者纳入了不同的膝关节假体RSA研究）被用于测试该系统的可重复性和准确性。X线片以600 dpi分辨率和256级灰度进行扫描。在PC显示器上识别出标记物后，计算机程序使用边缘检测软件算法根据标记物的轮廓计算X线片中每个钽标记物的中心。研究表明，基于计算机的图像分析可用于RSA检查。在RSA中使用图像处理软件的优点是，标记物位置的确定是客观的，并且无需在实际测量前对X线片上的所有标记物进行系统的人工识别。
