Low-Complexity Based Modified Image Super-Resolution Scheme by the Design of Dyadic Integer Coefficient Based Wavelet Filters

This paper presents a low-complexity based modified image super-resolution scheme based on the wavelet coefficients soft-thresholding .The design this scheme is based on  a particular class of dyadic-integer-coefficient based wavelet filters (DICWFs) which is formulated from the design of a half-band polynomial. To design integer-coefficient based half-band polynomial we used the splitting approach. Next, factorization is done for this designed half-band polynomial and assigned specific number of vanishing moments and ischolar_mains to achieve the dyadic-integer coefficients low-pass analysis and synthesis filters to reduce the hardware complexity. The discrete wavelet transform (DWT) obtained from DICWF is applied on the low-resolution image to obtain the high frequency sub-bands. These high frequency sub-bands and the original low-resolution image are then interpolated to enhance the resolution. Next, stationary wavelet transform (SWT) that are obtained using DICWFs is employed to minimize the loss due to the use of DWT. In addition, wavelet coefficients soft-thresholding scheme is used on these estimated high-frequency sub-bands in order to reduce the spatial domain noise. These sub-bands are combined together by inverse discrete wavelet transform obtained from DICWF to generate a high-resolution image. The proposed approach is validated based on quality metrics of existing filter banks and proposed filter banks.