Feature preserving image interpolation is an active area in image processingfield. In this paper a new direct edge directed image super-resolutionalgorithm based on structure tensors is proposed. Using an isotropic Gaussianfilter, the structure tensor at each pixel of the input image is computed andthe pixels are classified to three distinct classes; uniform region, cornersand edges, according to the eigenvalues of the structure tensor. Due toapplication of the isotropic Gaussian filter, the classification is robust tonoise presented in image. Based on the tangent eigenvector of the structuretensor, the edge direction is determined and used for interpolation along theedges. In comparison to some previous edge directed image interpolationmethods, the proposed method achieves higher quality in both subjective andobjective aspects. Also the proposed method outperforms previous methods incase of noisy and JPEG compressed images. Furthermore, without the need foroptimization in the process, the algorithm can achieve higher speed.
展开▼
