Wavelet Based Super-Resolution Approach with Improvement Using Sparse Representation


Author (s):

  1. Dr. D. D. Vyas, Darshan Institute of Engineering & Technology, Rajkot, Gujarat, INDIA
  2. Dr. D.R.Bhojani, Darshan Institute of Engineering & Technology, Rajkot, Gujarat, INDIA
  3. Kalpesh J. Kothari, Darshan Institute of Engineering & Technology, Rajkot, Gujarat, INDIA, kalpesh.kothari@darshan.ac.in

Abstract:

In the proposed super resolution reconstructed scheme the construction of high frequency subband images are done by using Lifting Wavelet Transform of the Luminance component of LR image. To enhance an intermediate stage by using SWT is introduced. In order to decompose luminance component of an input LR image into different subband images LWT is applied. Then the interpolations of high frequency subband images are done by using bicubic interpolation. The modifications in estimated high frequency subband images are being done by using high frequency subband obtained using SWT. Then to preserve the more information sparse representation is applied to Luminance component of LR image. Then sparesed image and modified high frequency subband images are combined to generate a Luminance component of high resolution image by using inverse SWT (ISWT). Then the Cb and Cr component of LR image is interpolated using bicubic interpolation. And at last the YCbCr to RGB conversion is done and High Resolution Colour Image can be generated. The quantitative and visual results are showing the superiority of the proposed technique over the conventional and state-of-art image resolution enhancement techniques.

No of downloads: 715