【摘要】 目的　旨在研究用于 X线头影测量分析的各标志点的计算机自动识别与定位方法 ,建立计算机自动定点的 X线头影测量分析系统 (CACAS)。方法　应用现代计算机图像处理与模式识别的相关技术 ,其中包括中值滤波、直方图均衡、基于小波变换的多尺度分析、Canny边缘检测、基于知识的边缘跟踪及可变模板匹配技术等 ,改善 X线头影图像质量 ,提取软硬组织边缘轮廓 ,自动定位识别标志点。结果　 CACAS系统成功实现了 17个标志点的自动识别与定位及面下 1/3轮廓描迹。结论　本研究建立 CACAS系统所运用的技术方法是有效可靠的。更多还原

【Abstract】 Objective To exploit the technique for automatic identification of landmarks and develop the computerized automatic cephalometric analysis system(CACAS) to identify the landmarks by computer instead of manual operation. Methods The techniques related to computer digital image processing and pattern recognition such as Median filtering, Histogram equalization, Canny edge detection, Wavelet transformation, Knowledge based edge tracing and changeable template were introduced to develop the computerized automatic cephalometeric analysis system(CACAS). The whole function of CACAS were realized by three steps as followed: firstly, image preprocessing including moving noise and image enhancement by applying Median filtering and Histogram equalization; secondly, extracting the soft-hand tissue edge of image by applying canny edge detection wavelet transformation, Knowdege based edge tracing and changeable template techniques; finally, landmark identification by identifying its geometric characters.Result Seventeen landmarks were successfully identified and low 1/3 face profile werre outlined by CACAS. Conclusion The techniques applied to develop the CACAS were effective and reliable. Especially the identification of landmarks in extra-edge of image is more precise than manual operation. But the landmarks of inner edge of image were not identified by computer at present. The study must be continued further.更多还原