摘要
对脑部损伤CT图像目标点的定位,能够有效提高医学上对脑损伤的诊断效率。对脑部损伤CT图像目标点的定位,需要获取脑部损伤标记边缘点的集合,提取其连通区域目标点,完成图像目标点的定位。传统方法在平滑后的梯度图像中提取标记,给出标记阈值,但忽略了提取边缘点连通区域目标点,导致定位精度较低。提出基于模糊聚类的医学二维脑部损伤CT图像目标点定位方法。该方法先通过改进的空域滤波器对含有噪声的二维脑部损伤CT图像进行滤波去噪,将CT图像变换成通用图像,促使输入的图像具有相同的直方图灰度分布状态,对每个分块区域进行分形维数计算,获取脑部损伤标记边缘点的集合,提取其连通区域目标点,完成对医学二维脑部损伤CT图像目标点定位。仿真证明,所提方法可以有效、准确地自动定位脑部损伤CT图像的损伤部位,为临床医生提供可靠的诊断依据。
An object point location method for 2D medical CT image of brain injury is proposed based on fuzzy clustering. Firstly, the 2D CT image of brain injury with noise is filtered and denoised through improved spatial filter, and transformed into general image, then the input image is made to have same gray histogram distribution state. The fractal dimension on each block region is calculated, and the set of marginal point in brain injury is obtained. In ad- dition, the target point in connected region is extracted. Thus, the object point location of 2D medical CT image of brain injury is completed. Simulation result shows that the proposed method can effectively and accurately locate in- jured region of CT image of brain injury. It can provide reliable diagnosis for clinicians.
作者
王哲
张钰
王勇
WANG Zhe;ZHANG Yu;WANG Yong(Institute of clinical, Hebei Medical University, Shijiazhuang Hebei 050000, China;Second Hospital, Hebei Medical University, Shijiazhuang Hebei 050000, China)
出处
《计算机仿真》
北大核心
2017年第11期228-231,共4页
Computer Simulation
