Speeding up edge guided image interpolation using the integral image technique

Abstract

Image interpolation and super-resolution are subjects of great interest in image processing. The main task behind any image interpolation method is to obtain superior highresolution images with improved visual quality, and recuperate rich texture and sharp edges using minimal calculation. In this sense, the aim of this paper is to present a new method which is effective in view of high resolution quality, processing speed, and economic performance by reducing the amount of calculations required to perform image edge directed image interpolation. This can be attained by operating in an integrated manner image interpolation using 4-directional image gradient information which defines differences in intensities between contiguous pixels. The proposed method is implemented and tested over several gray images, and compared to many interpolation methods in the state of the art. It has been demonstrated to be able to obtain faithfully better results than the habitual interpolation methods in terms of image quality metrics (PSNR, SSIM, FSIM) as well as the the perceptual visual appearance of the reconstructed images.