Super resolution enhancement of medical image using quaternion wavelet transform with SVD

Abstract

In this paper, a novel resolution enhancement approach based on Quaternion wavelet transform (QWT) with singular value decomposition (SVD) is proposed. The technique decomposes the input image into sixteen frequency sub bands by using QWT. The singular values of the low-low sub band image are estimated and the high frequency sub bands are interpolated using Lanczos interpolation. Finally, a contrast enhanced super resolution image is obtained by combining the interpolated high frequency sub bands and contrast enhanced image by inverse QWT. The visual and quantitative results show that the proposed QWT-SVD method clearly outperforms the bilinear, bicubic, DWT-bicubic, DTCWT-NLM-SVD with better edge preservation.