颅盖骨骨折CT图像的计算机自动分析研究  

A Calvarium Fractue Analyzer Using CT Image

作　　者：张春玲[1] 宋明军[2] 杜海涛[1] 宋吾力[1] 魏树刚 

机构地区：[1]泰山医学院,山东泰安271000 [2]泰安市第四人民医院,山东泰安271000 [3]山东泰安荣军医院,山东泰安271000

出　　处：《中国医学物理学杂志》2013年第6期4537-4542,共6页Chinese Journal of Medical Physics

基　　金：山东省教育厅项目(J11LF76)

摘　　要：目的:寻找一种方法,并设计程序,使计算机能够通过运行程序对从第三脑室上部层面到大脑皮层下部层面的CT图像进行自动分析,由此来判断图像中的颅盖骨是否存在骨折,从而为扩展CAD技术的应用做探索性研究。方法:对于第三脑室上部层面到大脑皮层下部层面的CT图像,首先进行预处理,去除图像中的噪声和头架;然后对预处理后的图像用阈值分割法进行分割,把颅盖骨从图像中分割出来;然后对得到的图像进行修正处理,使颅骨骨缝和分割时造成的孤立点对最后判断的影响尽量降低;然后对修正处理后的颅盖骨图像进行边缘提取;最后根据提取的颅盖骨边缘判断颅盖骨是否存在骨折并给出判断结果。结果:计算机运行根据所找到的方法编制的程序,测试了已明确诊断的61幅从第三脑室上部层面到大脑皮层下部层面的CT图像,结果有53幅判断正确,正确率为86.89%。结论:本研究找到的数据挖掘方法,通过程序实现,可以对从第三脑室上部层面到大脑皮层下部层面的CT图像进行自动分析,由此判断图像中的颅盖骨是否存在骨折。Objective： To study a Analyzer that can make computer could analyze CT images of one＇s skull from top of third ventricle to underlayer of pallium automatic, and judge whether having fracture in image. Methods： The analyzer can reduce noise of CT images of one＇s skull from top of third ventricle to underlayer of pallium at first, and then get the calvarium＇s image by segmentation and turn the image into binary image at the same time. The next work is amending the image by dilation and wiping off isolated point which could reduce the influence of skull slot and isolated point. And then the analyzer should detect the edge of calvarium from amended image. Finally analyzer could judge the calvarium whether having fracture. Results： Analyzer could tightly identify 52 images from 61 CT images of one＇s skull from top of third ventricle to underlayer of pallium （28 images with fracture, 33 images without fracture）, and the right ratio is 86.89%. Conclusions： The analyzer could analyze the CT images of one＇s skull from top of third ventricle to underlayer of pallium automaticly and then identify which image has fracture.

关 键 词：颅盖骨 骨折 计算机 自动分析 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]