An efficient approach for structure-texture image decomposition with edge-preservation

Abstract

The decomposition of an image into its structural part and texture part is a difficult task and has application in various areas such as medical imaging and image segmentation, etc. Various methods have been proposed to decompose an image into its structure part containing large scale variation in pixel intensity and texture part containing small scale variation in pixel intensity. Unfortunately almost all available methods in the literature have a common drawback that they remove or blur the fine geometric edges while removing the texture, as both geometrical edges and texture both belongs to the high frequency domain. In this work, we propose a new approach to deal with this drawback. The proposed approach effectively remove the texture without affecting the geometrical edges. Also we compare the experimental results with state-of-the-art methods to show the efficiency and effectiveness of proposed approach.