Regional Varying Image Super-Resolution

In this paper a new method for single image super-resolution using some high resolution images is proposed. It is assumed that each high resolution image shows a region of the low resolution image, with some differences about illumination or view point. These differences may be due to taking photos in different times, camera motion or unequal zooming. In the proposed method each high resolution image is mapped to a resized version of the given low resolution image using homography matrix and RANSAC method, which are known in computer vision context. The mapped image is fused with the LR image for producing a synthesized image. The mentioned method is repeated for each of the HR images. The resulting image has higher resolution only on its regions corresponding to HR images. The experimental results show the superior performance of the proposed method against some other methods in term of final perceived quality.