A Preconditioning Iterative Algorithm Based on Signal and Noise Separation
Preconditioning Landweber iterative method is an image deconvolution algorithm. The preconditioning strategy aims at accelerating the convergence speed. Unfortunately, it expands the transmission of noise simultaneously and leads to the decreasing accuracy of image restoration. The pre conditioning strategies tend to accelerate the convergence speed while also expanding the transmission of noise, which reduces the accuracy of image restoration. In this paper, based on signal and noise separation which can separate the signal domain and noise domain, we propose a preconditioning Landweber iterative algorithm which only improves the convergence speed in the signal domain, a preconditioning Landweber iterative algorithm expediting the convergence only in the signal domain is put forward. As a result, we can still improve the restoration accuracy of results at the same time of speeding up convergences.
Image deconvolution ill-posedness Landweber iterative method preconditions signal and noise separation
Shengchun Wang Yaping Huang
Department of Computer Science & Technology Beijing Jiaotong University Beijing, China
国际会议
成都
英文
901-905
2010-12-17（万方平台首次上网日期，不代表论文的发表时间）