Enhancing Video Deblurring using Efficient Fourier Aggregation

Abstract

Video Deblurring is a process of removing blur from all the video frames and achieving the required level of smoothness. Numerous recent approaches attempt to remove image blur due to camera shake,either with one or multiple input images, by explicitly solving an inverse and inherently ill-posed deconvolution problem.An efficient video deblurring system to handle the blurs due to shaky camera and complex motion blurs due to moving objects has been proposed.The proposed algorithm is strikingly simple: it performs a weighted average in the Fourier domain, with weights depending on the Fourier spectrum magnitude. The method can be seen as a generalization of the align and average procedure, with a weighted average, motivated by hand-shake physiology and theoretically supported, taking place in the Fourier domain. The method�s rationale is that camera shake has a random nature, and therefore, each image in the burst is generally blurred differently.The proposed system has effectively deblurred the video and results showed that the reconstructed video is sharper and less noisy than the original ones.The proposed Fourier Burst Accumulation algorithm produced similar or better results than the state-of-the-art multi-image deconvolution while being significantly faster and with lower memory footprint.The method is robust to moving objects as it acquired the consistent registration scheme.