基于特征点抽取的闭合曲线快速检测法在颅盖骨图像中的应用

目的:探讨基于特征点抽取的闭合曲线快速检测法在颅盖骨图像中的应用效果。方法:将颅盖骨图像提取出来并进行细化处理后,从图论的观点出发,提取图中的端点、歧点和叉点,根据Euler公式计算图像中的边数和顶点数,由此判断颅盖骨是否闭合。结果:测试已明确诊断的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56幅判断正确,正确率为91.8%。结论:实验表明,此方法可以用来判断颅盖骨的闭合情况。

Application of Fast Detection of Closed Curve in Image Based on Feature Point Extraction to Calvarium Image

Objective： To study the application of fast detection of closed curve in image based on feature point extraction to calvarium image. Methods： A thinned image of calvarium is viewed as a graph, and feature points such as ending, bifurcation and crossover are extracted as the vertexes. Based on Euler formula to count the number of vertexes and sides and analyze their quantitative relationship, whether fracture or not in the image. Results： The method could rightly identify 56 images from 61 CT images of one＇s skull from top of third ventricle to underlayer of pallium, and the right ratio is 91.8%. Conclusion： The method could identify which image has fracture.