VDN: Variant‐depth network for motion deblurring

Abstract

Motion deblurring is a challenging task in vision and graphics. Recent researches aim to deblur by using multiple sub-networks with multi-scale or multi-patch inputs. However, scaling or splitting operations on input images inevitably loses the spatial details of the images. Meanwhile, their models are usually complex and computationally expensive. To address these problems, we propose a novel variant-depth scheme. In particular, we utilize the multiple variant-depth sub-networks with scale-invariant inputs to combine into a variant-depth network (VDN). In our design, different levels of sub-networks accomplish progressive deblurring effects without transforming the inputs, thereby effectively reducing the computational complexity of the model. Extensive experiments have shown that our VDN outperforms the state-of-the-art motion deblurring methods while maintaining a lower computational cost.

Multi-Stage Attentive Network for Motion Deblurring via Binary Cross-Entropy Loss

“…Notably, in comparison with our previous approach [ 23 ], this work has the following main differences. (1) We insert a self-attention module with group convolution between the encoder and decoder based on the sub-network structure of the method [ 23 ].…”

“…Notably, in comparison with our previous approach [ 23 ], this work has the following main differences. (1) We insert a self-attention module with group convolution between the encoder and decoder based on the sub-network structure of the method [ 23 ]. That is mainly inspired by [ 30 , 31 ], enhancing the adaptability of the model by computing the correlation of relevant positions to obtain the most useful and important features and recalibrate these features.…”

“…These end-to-end methods do not need to estimate blur kernels, rather only the mapping learned between blurred images and sharp images, so as to directly generate predicted restored images from the blurred images. In particular, Nah et al [ 10 ] proposed a multi-scale deblurring neural network based on multi-stage networks (MSNs) for directly estimating latent sharp images from the blurred images, and other methods [ 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 ] further advance MSN models for deblurring. Moreover, generative adversarial network (GAN)-based methods [ 25 , 26 , 27 ] have also been used to achieve end-to-end image deblurring.…”

Precision in Pursuit: A Multi-Consistency Joint Approach for Infrared Anti-UAV Tracking

“…Yet, compared to RGB images, infrared imagery usually possesses lower resolutions and single-channel representations, implying a more constrained source of target-related information [7]. Additionally, due to UAVs' irregular motion patterns [8] and camera rotations, even in adjacent frames, target discontinuities may manifest, as illustrated in Fig. 1.…”

Nighttime Driver Behavior Prediction Using Taillight Signal Recognition via CNN-SVM Classifier

“…For example, Chen et al [13] proposed a self-calibrated pyramid aggregation network that uses a decoder-encoder architecture to remove haze in images. In [14], a solution to remove the blur and with the aim of reducing the computational complexity is provided. It should be noted that these items are designed on a caseby-case basis and only to solve a particular challenge and cannot be replaced with each other.…”