Online Video Deblurring via Dynamic Temporal Blending Network

By Tae Hyun Kim, Kyoung Mu Lee, Bernhard Schölkopf, Michael Hirsch. 2017

176 citations recorded by the source corpus.

State-of-the-art video deblurring methods are capable of removing non-uniform blur caused by unwanted camera shake and/or object motion in dynamic scenes. However, most existing methods are based on batch processing and thus need access to all recorded frames, rendering them computationally demanding and time-consuming and thus limiting their practical use. In contrast, we propose an online (sequential) video deblurring method based on a spatio-temporal recurrent network that allows for realtime performance. In particular, we introduce a novel architecture which extends the receptive field while keeping the overall size of the network small to enable fast execution. In doing so, our network is able to remove even large blur caused by strong camera shake and/or fast moving objects. Furthermore, we propose a novel network layer that enforces temporal consistency between consecutive frames by dynamic temporal blending which compares and adap- tively (at test time) shares features obtained at different time steps. We show the superiority of the proposed method in an extensive experimental evaluation.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Artificial intelligence in the creative industries: a review (2020, background): “A Spatio-temporal recurrent network exploiting a dynamic temporal blending network is proposed in [115].”
Adversarial Spatio-Temporal Learning for Video Deblurring (2018, background): “[28], DTBNet [50], DBLRNet (single), DBLRNet (multi), DBLRNet and DBLRGAN.”
Neural Blind Deconvolution Using Deep Priors (2019, background): “Besides, by exploiting the temporal information between adjacent frames, deep networks have also been applied to video motion deblurring [10,27,31].”
NTIRE 2019 Challenge on Video Deblurring and Super-Resolution: Dataset and Study (2019, background): “[13] employs a temporal blending module in a fast RNN architecture.”
Deep Image Deblurring: A Survey (2022, methodology): “…a selective parameter sharing scheme, which leads to improved performance over SRN. Recurrent layers canalsobeusedtoextract temporal information from neighboring frames in videos (Hyun Kim et al. 2017; Lumentut et al. 2019; Nah et al. 2019; Park et al. 2020; Zhong et al. 2020; Zhou et al. 2019).”
Scale-Recurrent Network for Deep Image Deblurring (2018, background): “More recent work started to use end-to-end trainable networks for deblurring images [25] and videos [18, 33, 37].”
Spatio-Temporal Filter Adaptive Network for Video Deblurring (2019, methodology): “Kim et al .”