骨龄X射线图像的预处理和改进k余弦的指骨关键点定位

背景：在骨龄自动化评价过程中,二值化阈值的选取直接关系到图像二值化效果以及对指骨关键点的准确定位。目的：解决骨龄X射线图像二值化阈值选取问题,以及k余弦算法定位指骨关键点速度过慢的弊端,从而达到快速准确定位指骨关键点的目的。方法：采用图像预处理技术,提出基于OSTU阈值的二次阈值算法,用该方法获取的阈值来二值化骨龄X射线图像;针对k余弦算法定位指骨关键点速度很慢的缺点,提出一种改进的k余弦算法来快速准确的定位指骨关键点。结果与结论：用改进后的k余弦算法来定位63例指骨关键点的总成功率在95%以上,此法运行速度比k余弦算法快20倍以上,其准确性和速度都达到比较满意的效果。

Hand radiographs preprocessing and improved k-cosine algorithm in location method for phalangeal key points

BACKGROUND：During the bone age assessment,binarization threshold selection is one of the most troubled factors,which is directly related to two-value image and accuracy location of the phalanx key points.OBJECTIVE：In order to solve the problems for binarization threshold selection and overcome the too slow speed for locating the key points of phalanx,to reach fast and accurate location for them.METHODS：In this paper,image preprocessing technique is presented and the second threshold algorithm on the base of OSTU threshold algorithm is also put forward respectively in the analysis of the skeletal age X-ray pictures.And the threshold obtained through this method is used to get binarization from the skeletal age X-ray pictures.Since k-cosine algorithm is very slow to locate the phalanx key points,an improved k-cosine algorithm is put forward to locate the phalanx key points,which is fast and accurate.RESULTS AND CONCLUSION：The experimental results show that the proposed method used to locate phalanx key points can reach satisfactory results in accuracy and speed.