HyeSuk Kim, Gihong Kim, Gueesang Lee, June-Young Chang, Hanjin Cho
{"title":"Anisotropic Diffusion for Preservation of Line-edges","authors":"HyeSuk Kim, Gihong Kim, Gueesang Lee, June-Young Chang, Hanjin Cho","doi":"10.1109/ISSPIT.2008.4775703","DOIUrl":null,"url":null,"abstract":"In existing approaches, diffusion is performed in four directions (North, South, East, West) without specific conditions. Therefore, these methods have shortcomings of distorted with the existence of impulse noises. In this paper, a new anisotropic diffusion based on directions of line-edges is proposed to enhance preservation of line-edges together with removal of noises. In the proposed method, an edge detection mask is used to find the direction of a line-edge. As a result, when the magnitude of edge detection is large enough, there exists a line-edge. In the case of a line-edge, the weight of diffusion is selected adaptively according to the direction of the line-edge. The diffusion is based on 8-directions diffusion with emphasis on the line-edge direction. Experimental results show that the proposed method can eliminate noise while preserving contour of line-edges.","PeriodicalId":213756,"journal":{"name":"2008 IEEE International Symposium on Signal Processing and Information Technology","volume":"48 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2008-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2008 IEEE International Symposium on Signal Processing and Information Technology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ISSPIT.2008.4775703","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
In existing approaches, diffusion is performed in four directions (North, South, East, West) without specific conditions. Therefore, these methods have shortcomings of distorted with the existence of impulse noises. In this paper, a new anisotropic diffusion based on directions of line-edges is proposed to enhance preservation of line-edges together with removal of noises. In the proposed method, an edge detection mask is used to find the direction of a line-edge. As a result, when the magnitude of edge detection is large enough, there exists a line-edge. In the case of a line-edge, the weight of diffusion is selected adaptively according to the direction of the line-edge. The diffusion is based on 8-directions diffusion with emphasis on the line-edge direction. Experimental results show that the proposed method can eliminate noise while preserving contour of line-edges.