基于大周边感受野的仿生图像滤波器  

A bionic image filter using extended surround receptive field

作　　者：杜晓凤[1] 李翠华[1] 李玲玲[2] 

机构地区：[1]厦门大学信息科学与技术学院,福建厦门361005 [2]郑州航空工业管理学院计算机科学与应用系,河南郑州450015

出　　处：《华中科技大学学报（自然科学版）》2009年第11期78-81,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家重点基础研究发展计划资助项目(2007CB311005);福建省自然科学基金资助项目(A0710020);河南省教育厅自然科学基础研究计划资助项目(2007510023);河南省高等学校青年骨干教师资助计划项目;厦门大学985二期信息创新平台资助项目

摘　　要：针对图像处理的噪声平滑和对比度增强问题,提出了一种仿生图像滤波器.滤波器模拟初级视皮层自适应机制和大周边感受野(包括经典感受野、经典外周与去抑制区)3个区域之间的作用机制;引入自适应形状的核函数模拟自适应方向选择机制以平滑噪声;进而利用3个不同尺度的核函数来模拟大周边感受野的3个区域,通过3个区域的加权组合调节边缘区域的对比度.对灰度图像的处理实验表明:这种仿生图像滤波器可以有效地平滑图像噪声,增强细节特征,使图像在视觉效果上得到明显改善.De-nosing and contrast enhancement in image processing were dealt. Then a novel bionic im age filter was proposed. The proposed filter simulated the spatially adaptive mechanism in primary visual cortex and the mechanism of the extended surround field （which included the classical receptive field, classical surround and dis inhibition area）. In order to smooth noises, shape adaption of kernel was introduced which simulated the spatially orientation-selective mechanism. Furthermore, three dif ferent bandwidth kernels were introduced to simulate these three areas of extended surround field and their weighted combination was used to adjust the contrast of edge regions. Experimental results on grey images show that this bionic filter is a fruitful method to smooth noise, enhance weak information and improve visual quality of image.

关 键 词：图像滤波器 图像增强 空间自适应 梯度结构张力 感受野 

分 类 号：TN911.73[电子电信—通信与信息系统]