Image interpolation using oriented wavelet transforms

Abstract

A novel directionally adaptive image interpolation method based on the oriented wavelet transforms is proposed. The method uses the wavelet transforms built along multiple di- rections (directionlets) to e±ciently capture directional features and to extract edge information from the low-resolution image. Then, the high-resolution image is synthesized using the extracted information to preserve sharpness of edges and texture. The novel interpolation algorithm outperforms the state-of-the-art methods in terms of both numeric and visual quality of the interpolated image.