模糊生物图像特征优化提取仿真研究  

Fuzzy Biological Image Feature Extraction Simulation Research

作　　者：路晓亚[1] 杜丽娟[1] 

机构地区：[1]商丘工学院信息与电子工程学院,河南商丘476000

出　　处：《计算机仿真》2017年第5期397-400,共4页Computer Simulation

基　　金：河南省教育厅科学技术研究重点项目(12B520047)

摘　　要：对模糊生物图像特征的优化提取,可更好的识别不同程度的模糊失真图像。对图像的特征进行提取,需要计算出图像局部像素点邻域内的梯度值,定义图像特征标志函数,完成对模糊图像的特征提取。传统方法提取不同属性的特征基图像,给出模糊生物图像特征目标函数,但忽略了求取图像像素点的梯度值。提出基于单层垂直上下文的模糊生物图像特征优化提取方法。该方法融合于模糊变换理论给出图像各区域模糊状态信息,计算出模糊生物图像局部可分性判据,融合于SIFT特征提取方法获取模糊生物图像特征的关键点,计算关键点对应的梯度方向,确定模糊生物图像局部中心像素是否与周围像素处在同一区域,计算出图像局部像素点邻域内的梯度值,定义图像特征标志函数,以此为依据完成对模糊生物图像特征优化提取。仿真证明,所提方法模糊生物图像特征提取精度高,提升了对模糊生物图像的识别效率。An optimization feature extraction method for blurred biological image is presented based on context of single layer perpendicular. Integrated with fuzzy transformation theory, blurred status information in each area of im- age is provided to work out local separable criterion of blurred biological image. Integrated with SIFT feature extrac- tion method, key point of image feature is obtained and gradient direction corresponding to key point is calculated. Moreover, it confirms whether local center pixel and around pixel are in the same area. Finally, gradient value in neighborhood of local pixel point of image is worked out and mark function of image feature is defined. Thus, the op- timization of feature extraction is completed. Simulation result shows that the proposed method has high extraction pre- cision. It improves the recognition efficiency.

关 键 词：模糊生物图像 特征 优化提取 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]