基于小波多尺度分析的X—线头影特征点提取

提出了一种基小波多尺度分析的X－线光头影特点自动提取方法，先进行多尺度分解，得到不同分辨率的图像；接着在低分辨率的图像上处理，用Canny边缘检测算子提取边缘；用基于知识的跟踪方法和可变模板法连接边缘，得到了骨骼线和软组织线；将轮廓线作为指导信息，在高分辨率图像上校正，得到高精度的轮廓线；根据特征点的几何定义，利用轮廓线的益信息、曲率等属性来提取征点，实验表明，对于质量较好的X－线图像，该方法具有较高定位精度，并且稳健性较好。

Characteristic Points Extraction of X-Ray Skull Image Based on Wavelet Multiscale Analysis

A wavelet based method was presented to automatically extract characteristic points of human skull. The processing procedure is as follows: the X ray images are decomposed by wavelet, and the multi scale images are acquired; on the low resolution images, Canny filter is used to get the edges of these images; a knowledge based tracking method and knowledge based deformable template is used to get the consistent contours. Based on the guidance of these contours and studying on the high resolution images, we can get the accurate contours. Lastly, the characteristic points can be extracted by the geometry definition of them and the position and curvature of each point on the contours. The experiments show that this method is accurate and robust for the preferably high quality X ray image.