非经典感受野的朝向和对比度特性在轮廓提取中的应用

对于普通的复杂背景图像的边缘检测方法,在检测出主体轮廓的同时也保留了背景的纹理,不利于检测出主体的轮廓.根据生物视觉感知的特性,模拟出复杂细胞感受野的响应函数,然后提出了一种基于朝向和对比度特性的非经典感受野抑制模型,从而得到了清晰的轮廓.仿真实验证明,该方法在提取物体轮廓的同时有效地抑制了背景的纹理,提高了轮廓提取的正确率.

Method of Non-classical Receptive Field Contour Extraction Based on Orientation and Contrast Property

For the ordinary edge detection method with complex background image,the texture of the background is reserved while the main outline is detected.So it is not conducive to detect the outlines of the subject.According to the characteristics of the biological visual perception,resp-onse function of complex cell receptive field has been simulated,then a non-classical receptive field inhibition model based on orientation and contrast characteristics is proposed.Simulation res-lts show that the method can extract the contour of the object and effectively inhibit the texture of the background at the same time,which improves the contour extraction accuracy rate.