【摘要】 目的 :实现头颅侧位片标志点的自动识别 ,减少人工定点带来的误差。方法 :确定标准模板并确定标准图像上标志点的位置 ,利用仿射变换将这些点变换到被测图像上作为初始位置 ,以缩小搜索范围 ;利用对标准模板的形变实现不同标志点的识别。结果 :实现软硬组织轮廓的提取和一些孤立点的识别 ,对依附于轮廓的点的精度较高。结论 :该方法能较好的识别头颅侧位片中常用的测量点。更多还原

【Abstract】 Objective:Automatically identifying landmarks of the Lateral Cephalogram . Method:Establishing standard templates and marking many standard landmarks .By affine transformation , the searching area of landmarks were be reduced . The standard templates were deformed due to different landmarks .Result:The outline of rigid and soft tissue were recognized and some isolated points were identified. An evaluation of this technique in comparison to the accuracy was presented . Most of landmarks were within 2mm and landmarks near!boundaries were within 1 mm .Conclusion:Landmarks of the Lateral Cephalogram can be identified very well .更多还原