On a compact non-extrapolating scheme for adaptive image interpolation

The motivation of this paper is to improve the results of ENO subcell resolution, that already obtains really good results but using extrapolation, as was stated by F. Arándiga, R. Donat and P. Mulet [Adaptive Interpolation of Images, Signal Processing 83(2) (2003) 459-464]. We present a new method for image interpolation which combines a new nonlinear cell-average interpolatory technique that uses a trigonometric mean and subcell resolution. The scheme uses a fully compact reconstruction and avoid any step of extrapolation. The experiments presented validate the theoretical results obtained.