摘要
背景:在骨龄自动化评定中,许多方法对骨骺兴趣区域的提取不够理想。目的:通过k余弦算法定位出手指骨的特征点,解决对骨骺兴趣区域的提取的困难和提取方法的局限性。方法:提取指骨兴趣区域,并用各向异性扩散方法进行滤波,然后根据骨龄指骨的图像数据进行统计,寻找出定位指骨中心点的方法,通过3次多项式对中心点线进行拟合,定位出指骨的中心轴,在此基础上提取出指骨骨骺兴趣区域。结果与结论:用该方法定位出的指骨中心轴比较真实的反映指骨骨架,在此基础上能提取出指骨骨骺兴趣区域,成功率>95%,并能运用到后续对指骨骨骺兴趣区域相关的评定中。
BACKGROUND:At present,several methods are dissatisfactory for extracting epiphyseal/metaphyseal region of interest(EMROIs) ,which leads to poor reorganization effects and assessment accuracy of bone age.OBJECTIVE:To overcome the difficulties of EMROI extraction and extraction method limitations based on k-cosine algorithm to local key points of the phalanx.METHODS:EMROI were extracted and filtered by anisotropic anisotropic diffusion.The statistics of phalanx images data were analyzed and found out the method for locating the central points of the phalanx.A third-order polynomial was used to fit the central axis line according to the central points,and epiphyseal ROIs were then extracted.RESULTS AND CONCLUSION:Great deals of experiments are showed that the central axis line located by this method is correct,which reflects the real skeleton of the phalanx.On this basis,EMROI can be extracted,and the success rate rises to 95%.And the method can be applied to evaluate EMROI in the following research.
出处
《中国组织工程研究与临床康复》
CAS
CSCD
北大核心
2011年第30期5535-5538,共4页
Journal of Clinical Rehabilitative Tissue Engineering Research
基金
重庆市市渝中区科技计划项目
课题名称"骨龄自动评定系统的关键技术研究及应用"~~