摘要
融合视皮层边界响应与显著性信息,提出一种图像轮廓检测新方法。利用高斯导函数获得初级视皮层(V1)的多方向梯度响应,同时基于感受野机制设置Leaky integrate-and-fire神经元网络的突触兴奋和抑制电流;通过神经元脉冲频率信息,获得视觉显著性信息;反馈并调控V1的视觉响应,得到图像轮廓。以Ru G图像库为例,与ISO等方法相比,该方法在综合评价指标上具有优势。基于视觉显著性信息的传递和处理过程,将为后续的图像分析和理解提供新思路。
By combining the visual cortical boundary response and saliency information,a new method of image contour detection is proposed. First,the derivative of Gaussian transform was introduced to get the gradient responses of primary visual cortex（ V1） in multiple directions. Next the synaptic excitatory and inhibitory currents were specified based on receptive field mechanism in the LIF-neuron network. Visual saliency information was obtained by coding the spike frequency. Finally,the responses of neurons in V1 were modulated by feedback to get the image contour. The pictures used in this experiment were selected from the rug library. The results demonstrated that the proposed method had a significantly better performance than the traditional methods including ISO. The method for transmission and processing of visual saliency information will provide new ideas for the image analysis and understanding.
出处
《计算机应用与软件》
2017年第10期232-235,287,共5页
Computer Applications and Software
基金
国家自然科学基金项目(61501154)
关键词
轮廓检测
显著性信息
初级视皮层
脉冲频率
Contour detection Saliency information Prmary visual cortex Spike frequency
