一种鲁棒的骨龄X射线平片自动轮廓提取方法

An robust automatic contour extraction method for skeletal bone age X-ray images

背景:骨龄X射线平片具有不均匀和复杂性,因而在骨龄自动评价的研究中,手掌轮廓提取的结果往往不理想。目的:采用计算机自动提取手掌轮廓,为骨龄自动评价中图像预处理阶段的研究奠定重要基础。为了解除骨龄评定带来的主观性和不确定性,提出用计算机进行自动评价。方法:在仔细分析骨龄X射线平片的基础上,提出了一种对图像背景的可行有效子采样点方案,并提出用二元三次线性回归方法来模拟图像背景,通过形态学以及二进制标记等一系列操作,最后成功提取出手掌轮廓。结果与结论:采用异常点移除和回归方法相结合来提取轮廓,采取固定阀值,不受阀值选取的困扰,并具有鲁棒性。大量的实验结果表明了该方法提取的手掌轮廓成功率在93%以上,能完全应用于骨龄自动识别的后续研究工作。

BACKGROUND：Due to the non-uniform and complexity of bone age X-rays images,the results of contour extraction in automatic skeletal bone age assessment are often unsatisfactory. OBJECTIVE：To extract palm contour automatically using computer to lay a foundation for image pretreatment of bone age assessment to eliminate subjectivity and uncertainty in bone age assessment. METHODS：Bone age X-ray film was analyzed,and a feasible and effective program on the background image sub-sampling was proposed. Two-dimensional third order linear regression method was proposed to simulate the background images. Using a series of operations including morphological and binary image labeling,the contour of the hand was successful extracted. RESULTS AND CONCLUSION：By combining outliers removal with regression,a fixed threshold was used,which avoids selection of appropriate threshold. Moreover,it is robust. Numerous studies have demonstrated that the success rate of hand contour extraction was over 93%,so this method can be fully used in automatic identification of bone age research.