1 April 2011 Midfrequency-based real-time blind image restoration via independent component analysis and genetic algorithms
Author Affiliations +
Abstract
Blind image restoration (BIR) is a useful tool in image processing but too time-consuming to be applied. In order to truly achieve real-time BIR, an ultrafast method for BIR is proposed based on a new concept called the midfrequency region (MR) of an image, which can be used to estimate the point spread function (PSF) and rapidly and accurately restore the image. Also, new MR-based Wiener filtering is derived to accomplish BIR more quickly. Then, using the PSFs of class G that fit numerous imaging systems, a new MR-based algorithm for real-time BIR is derived via independent component analysis and genetic algorithms. Experiments prove that the algorithm is quite effective, speedy, and stable and can truly accomplish real-time image deblurring in on-line systems.
© (2011) Society of Photo-Optical Instrumentation Engineers (SPIE)
Yihan Luo, Chengyu Fu, "Midfrequency-based real-time blind image restoration via independent component analysis and genetic algorithms," Optical Engineering 50(4), 047004 (1 April 2011). https://doi.org/10.1117/1.3567072 . Submission:
JOURNAL ARTICLE
11 PAGES


SHARE
Back to Top