Image interpolation based on adaptive edge-preserving algorithm
GU Yan-feng · Ha'erbin gongye daxue xuebao · 2005
A new edge-preserving algorithm based on linear prediction model is proposed in order to preserve edge information better. According to geometric duality between high-resolution ( HR) image and low-resolution (LR) image (i. e. the invariability of edge feature in different resolution) , the new method is used to compute the optimal interpolation parameter vector by means of substituting HR linear predict model for LR one. The validity of edge-preserving of the model is testified by concrete examples. Experiment results show that many merits such as adaptability, edge preserving and lower computational complexity are contained in this method.