基于GABOR滤波的大脑MRI图像边缘检测  

MRI Image Edge Detection Based on Gabor Filter

作　　者：王琼[1] 彭延军[1] 董焕河[1] 

机构地区：[1]山东科技大学信息科学与工程学院,山东青岛266510

出　　处：《系统仿真学报》2006年第z1期360-362,共3页Journal of System Simulation

基　　金：山东省教育厅计划项目(J05c10)

摘　　要：基于二维Gabor滤波提出了一种对大脑MRI图像进行边缘检测与提取的方法,通过分析、简化、分解Gabor函数,设计了Gabor函数的多个参数,包括中心频率φ,滤波方向θ,以及沿着x轴和y轴Gaussian包络的空间常量σ。在算法实现过程中,恰当选取了一组可以覆盖给定图象整个频率空间的滤波,用这组滤波分别提取图象的局部信息,然后按照一定的规则将滤波后提取出局部信息的多幅图像整合为一幅图像。实验结果表明,该方法相对于传统的Gabor滤波可明显的减少计算量,并可连续、有效的检测边缘信息,定位准确,检测效果明显,对噪声也显示出较好的鲁棒性。A new method of edge detection and extraction of Medical image using 2-D Gabor filters was proposed. By analyzing and simplifying the Gabor function , several parameters were designed including central frequency φ ,orientation θ and the Gaussian envelope σ along the x-axis and y-axis .In this algorithm, a set of Gabor filters that can cover the entire frequency space for the given image were tuned,which is used to obtain the local information of the image, and then an accurate edge information can be developed by integrating this set of the Gabor filter outputs. Results of experiment show that this method can reduce computational requirements than the traditional Gabor filters ,and detect the image edge efficiently and continuously ,and is robust to the additional noise.

关 键 词：边缘检测 GABOR滤波 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]