An adaptive algorithm for image resolution enhancement

We present a novel and fast algorithm to enhance the resolution of an image by an arbitrary factor which is based on adaptively selecting different interpolating filters depending on local image analysis. Local image regions are analyzed in terms of edges and texture. This local image analysis is then used to select appropriate filters. We show that our algorithm significantly outperforms the standard bicubic interpolation method and does not suffer from ringing and blurring artifacts. Furthermore, we show a variant of the algorithm which works very well for scaling of bilevel images.