基于双通路视觉系统的自适应轮廓检测模型  

Adaptive contour detection model based on dual-pathway visual system

作　　者：王宪保[1] 陈斌[1] 项圣 陈德富[1] 姚明海[1] WANG Xianbao;CHEN Bin;XIANG Sheng;CHEN Defu;YAO Minghai(College of Information Engineering,Zhejiang University of Technology,Hangzhou 310023)

机构地区：[1]浙江工业大学信息工程学院,杭州310023

出　　处：《高技术通讯》2024年第1期15-24,共10页Chinese High Technology Letters

基　　金：国家自然科学基金(61871350);浙江省基础公益研究计划(LGG19F030011)资助项目。

摘　　要：在轮廓检测领域,背景纹理的干扰容易造成轮廓提取不完整。针对这一问题,本文提出了一种基于双通路视觉系统的自适应轮廓检测模型。首先从皮层下通路的信息采集与评估过程出发,对图像整体的显著性进行评估,以此获得轮廓信息的可能性分布;然后采用自适应尺度的高斯导函数对经典视觉通路中感受野的动态特性进行模拟,加强了模型对轮廓细节的捕获;最后在外周抑制算法的基础上,结合像素的空间分布对所有边缘的稀疏性进行度量,更加准确地区分了轮廓和纹理边缘。实验结果表明,本文模型可以有效抑制背景纹理,提升轮廓连续性,具有较好的轮廓检测性能。In this paper,an adaptive contour detection model based on dual-pathway visual system is proposed to solve the problem of incomplete contour extraction due to the interference of background texture.First,the information acquisition and evaluation process of the subcortical pathway is used to evaluate the saliency of the image as a whole,so as to obtain the probability distribution of contour information.Then,the dynamic properties of the receptive field in the classical visual pathway are simulated using adaptively scaled Gaussian derivative functions to enhance the capture of contour details by the model.Finally,based on the surround inhibition algorithm,the sparsity of all edges is measured in conjunction with the spatial distribution of pixels,which allows for a more accurate distinction between contour and texture edges.The experimental results show that the model proposed in this paper can effectively inhibit the background texture,improve the contour continuity,and have better contour detection performance.

关 键 词：轮廓检测 视觉机制 显著评估 感受野 稀疏度量 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]