A Novel Image Upscaling Method Using High Order Error Sharpening

A novel sophisticated interpolation scheme for converting low resolution images to higher resolution images is presented. Because of prediction, some blurring is introduced in the edges of the high resolution image while performing up-sampling. As a result, one post-processing scheme is used to eliminate the blurring artifact. The proposed interpolation scheme makes two major contributions. First, the high frequency degradation due to interpolation is identified. The degraded edge is sharpened using a high order laplacian filter in the next step. For various natural images, the proposed interpolation scheme is tested. Experiment results show that the proposed interpolation scheme outperforms the other state-of-the-art interpolation methods.