Joint Motion Deblurring with Blurred/Noisy Image Pair

Abstract

Motion blurred images are widely existing when using a hand-held camera especially under the dim lighting conditions. Since edge information contained in the noisy image may be blurred by the motion blur, a blurred/noisy image pair captured under different exposure time can help to restore a sharp image. In the traditional deblurring methods based on blurred/noisy image pair, the deblurring process is in series with the denoising process, so that restoration result is sensitive to the denoised result. In this paper, we propose a robust algorithm to obtain the sharp image by fusing the blurred image and noisy image. By joint modeling the deblurring model and denoising model, the restoration result can be optimized via estimating the sharp image and blur kernel alternately in the proposed methods, and it is not sensitive to the denoised result benefited by the joint model. Experimental results demonstrated that the proposed method can achieve better performance compared with the state-of-the-art single image denoising methods, single image deblurring methods and blurred/noisy pair deblurring methods.

Combining Motion Compensation with Spatiotemporal Constraint for Video Deblurring

“…A straightforward idea for this problem is to apply existing single or multiple image deblurring methods to each blurry frame [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 ]. For now, there are many mature single image deblurring methods [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 ].…”

“…Considering that more information will be conducive to the deblurring process, some other methods make the deblurring problem more tractable by leveraging additional input and joint multiple blurry images [ 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 ] in video recovery using block match multi-frame motion estimation based on single pixel cameras [ 9 ]. Tan et al compared a blurry patch directly against the sharp candidates in spatial domain, in which the nearest neighbor matches could be recovered [ 10 ].…”

Image Restoration with Multiple Hard Constraints on Data-Fidelity to Blurred/Noisy Image Pair

“…In the final experiment, we demonstrate that the setting of the parameters on data-fidelity in our method are much easier than the existing methods using a blurred/noisy image pair in the restoration step [14], [15]. In the existing methods, image restoration is performed by minimizing the sum of some specific regularization term and two data-fidelity terms on a blurred/noisy image pair, which can be expressed as the following optimization problem:…”

“…The model assumes that the blurred image v 1 also contains noise (usually σ 1 < σ 2 ), which is a realistic setting as addressed in [14], [21].…”

“…In addition, if noise in the blurred image is not negligible, restoring a sharp image from it becomes difficult even when using the true kernel. Meanwhile, the methods [14], [15] exploit the information on both images in their image restoration step. In these methods, the image restoration problem is formulated as the minimization of a regularization term, reflecting prior information on the latent image, plus two data-fidelity terms, keeping the consistency to a blurred/noisy image pair, where the balance among these terms is controled by multiple weights.…”