Fused Restoration Filtering for Single-Image Blind Deblurring
Fused Restoration Filtering for Single-Image Blind Deblurring
Aftab Khan*, Yasir Khan and Saleem Iqbal
ABSTRACT
Restoration filters are used in the Blind Image Deconvolution (BID) method to extract the clear image from the blurred image. Traditional restoration filters require manual estimation of its parameters in the traditional restoration case. Due to this parameter adjustment, restoration filters require parameter estimation which itself is a tedious job. This study proposes a novel technique for using image fusion to dispose the restoration filter’s parameter tweaking processing. Image fusion is used to produce a single deblurred image by firstly generating multiple deblurring results through a restoration filter and onwards combining them. Consequently, image fusion produces a single high-quality deblurred image relinquishing the parameter tuning process and making deblurring independent of the restoration filter variability. Simulations were performed on various image datasets, including tests of different blur types and restoration filters. Both naturally blurred and artificially blurred images have been used to test the proposed method. The efficiency of fusion-based restoration filtering is demonstrated by the experimental findings. Results show that the fused restoration filters provide enhanced parameter estimates and high-quality deblurred images. The Wiener filter surpasses in performance from the Total Variation (TV) and Richardson-Lucy (RL) filters by 3 and 98 percent respectively. When compared to the Weiner filter and TV deblurring in the case of naturally blurred images, the Richardson-Lucy deblurred images are 52 and 41 percent higher in terms of quality. The process can further benefit from the use of super-resolution techniques. This study’s fusion-based approach not only advances the field of blind image deblurring, but it also has great potential to improve image quality and accuracy across a range of domains, including surveillance systems, medical imaging, and satellite imagery, where precise visuals are essential for making important decisions.
To share on other social networks, click on any share button. What are these?
