Deep Dynamic Scene Deblurring from Optical Flow

TL;DR

This paper introduces a novel end-to-end deep learning framework that leverages multi-scale optical flow estimation and recurrent neural networks to effectively remove dynamic scene blur, outperforming existing methods in accuracy and efficiency.

Contribution

The paper proposes a multi-scale spatially variant RNN approach using optical flow for dynamic scene deblurring, enabling end-to-end training and improved performance.

Findings

Outperforms state-of-the-art algorithms in accuracy.

Achieves faster processing speeds.

Uses less model parameters.

Abstract

Deblurring can not only provide visually more pleasant pictures and make photography more convenient, but also can improve the performance of objection detection as well as tracking. However, removing dynamic scene blur from images is a non-trivial task as it is difficult to model the non-uniform blur mathematically. Several methods first use single or multiple images to estimate optical flow (which is treated as an approximation of blur kernels) and then adopt non-blind deblurring algorithms to reconstruct the sharp images. However, these methods cannot be trained in an end-to-end manner and are usually computationally expensive. In this paper, we explore optical flow to remove dynamic scene blur by using the multi-scale spatially variant recurrent neural network (RNN). We utilize FlowNets to estimate optical flow from two consecutive images in different scales. The estimated optical flow provides the RNN weights in different scales so that the weights can better help RNNs to remove blur in the feature spaces. Finally, we develop a convolutional neural network (CNN) to restore the sharp images from the deblurred features. Both quantitative and qualitative evaluations on the benchmark datasets demonstrate that the proposed method performs favorably against state-of-the-art algorithms in terms of accuracy, speed, and model size.