Fast image interpolation using directional inverse distance weighting for real-time applications

Abstract A novel simple image interpolation method with adaptive weights, which is motivated by the inverse-distance weighting (IDW) method, is proposed. A new distance is defined to implement the IDW-based algorithm. The weights corresponding to four diagonal pixels are computed in their own diagonal directions. And in order to make the algorithm robust to fine structures and noises, the weights are calculated in local windows. The new approach can be implemented efficiently and gives weights adaptively according to the local image structures. Experimental results demonstrate that the proposed method can generate visually pleasant interpolated images with high peak signal-to-noise ratios (PSNR) values in real time.