GRADIENT COLOR TENSOR BASED APPROACH FOR SPECTRAL MATTING
21 February 2013
Natural image matting is an essential task in image editing applications that has a fundamental role in TV and cinema industries. Image matting aims to extract foreground objects from a given image in a fuzzy mode where, for each pixel, an alpha degree of its belonging to the foreground or to the background is determined. One of the major state-of-the-art methods in this field is spectral matting. This latter automatically computes fuzzy matting components by using the smallest eigenvectors of a defined Laplacian matrix that is generated from affinities computation between adjacent pixels in an input image. The results obtained by such approach are coarsely related to the ability of defining affinity matrix that it should be able to well separate between different pixel clusters. To accomplish better matting and get better results, we propose a new spectral matting approach. We use a color tensor gradient of multispectral images in order to enhance the affinity computation process. Experiments through both quantitative measurements and visual results show that the performance and the efficiency of our proposed method improve the foreground extraction process compared to the original approach.
