Image super resolution using fusion of images

In this paper, an approach for image super resolution based on the fusion of images is proposed. The input image is interpolated and then it is subjected to direction lifting scheme to decompose it into subbands. Parallel to this, the image subbands are obtained by applying stationary wavelet transform (SWT) on the input image. The SWT transform is applied to the image for edge enhancement. The corresponding subbands obtained by applying SWT and direction lifting scheme are fused to obtain one output image. The second output image is obtained using the conventional interpolation method. The third output image is obtained by applying the midpoint algorithm which helps in preserving the average intensity of the image. All output images are fused together to generate a super resolution image of an input image. Peak signal-to-noise ratio is taken as a para to evaluate the quality of the super resolution images. Experimental results show that the proposed approach produces better results in comparison to various existing algorithms.