Adaptive contour detection model based on dual-pathway visual system
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 re-ceptive 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 spar-sity 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.
国家科技期刊平台
NSTL
万方数据
维普
