腹部CT图像肾脏自动分割方法研究

Study on automatic kidneys segmentation from abdominal CT images

提出一种基于先验知识的肾脏自动分割算法。首先利用肾脏位置的先验知识以肋骨和脊椎骨为界标选取感兴趣的肾脏区域,然后使用OTSU阈值分割法去除感兴趣的肾脏区域中灰度值较低的肌肉和脂肪组织,最后对不同类型的肾脏图像使用不同的方法进行分割。实验结果表明,该方法能够高效准确地分割出腹部CT图像中的肾脏,研究成果在临床肾脏疾病诊断上有一定的实用价值。

This paper proposed an automatic kidney segmentation algorithm based on some prior knowledge about kidney.Firstly,used the location of ribs and spine as the landmark for extraction of interested kidney regions.Secondly,used the OTSU thresholding approach in interested kidney regions to eliminate muscle and body fat whose gray level was low.Finally,used different methods to segment different kinds of kidney images.The experiment results show that the proposed method can efficiently and accurately segment the kidney in abdominal CT images,and has a certain degree of practical value on clinical diagnosis of kidney disease.