提高CT图像手骨模型重建精度的方法

A Method of Increasing Precision for Hand Bone Models Reconstruction of CT Images

为提高在体骨骼模型的3D重建精度,对CT扫描方式和利用CT图像重建骨骼模型的方法进行研究,提出一种基于灰度特征搜索匹配点的图像层间插值算法.该算法根据插值点相邻层邻近点的灰度值特征选择参考点;以参考点为依据在另一相邻层图像某区域内搜索与参考点灰度相近并且距离较小的匹配点;利用参考点和匹配点灰度值以及距离关系计算插值点灰度值.对人体9块手骨进行模型重建,将原始CT图像重建模型和插值优化后的重建模型分别与解剖模型进行比较的结果表明,优化前与优化后平均正偏差均值由0.30 mm减小到0.11 mm,平均负偏差均值由0.27 mm减小到0.09 mm.

In order to improve the precision of3D bone model reconstruction,the CT scanning mode and thebone model reconstruction method using CT images were studied,an inter-slice image interpolation algorithm based on gray feature searching matching point was proposed.The algorithm selects the reference point according to the gray value of the neighboring point of a layer adjacent to the interpolation point;searching for a matching point similar to the reference point gray and the smaller distance in the area of another adjacent layer image based on the reference point;The reference point,the matching point gray value and the distance relationship are used to calculate the interpolation point gray value.Nine hand bone models of human body were reconstructed,and the reconstructed models before and after interpolation were compared with the anatomical models respectively.Results from this study indicate that the mean deviation of average deviation was reduced from0.30mm to0.11mm,and the mean negative deviation was decreased from0.27mm to0.09mm.