Abstract:Recent research on deep convolutional neural networks (CNNs) has provided a significant performance boost on efficient super-resolution (SR) tasks by trading off the performance and applicability. However, most existing methods focus on subtracting feature processing consumption to reduce the parameters and calculations without refining the immediate features, which leads to inadequate information in the restoration. In this paper, we propose a lightweight network termed DDistill-SR, which significantly improves the SR quality by capturing and reusing more helpful information in a static-dynamic feature distillation manner. Specifically, we propose a plug-in reparameterized dynamic unit (RDU) to promote the performance and inference cost trade-off. During the training phase, the RDU learns to linearly combine multiple reparameterizable blocks by analyzing varied input statistics to enhance layer-level representation. In the inference phase, the RDU is equally converted to simple dynamic convolutions that explicitly capture robust dynamic and static feature maps. Then, the information distillation block is constructed by several RDUs to enforce hierarchical refinement and selective fusion of spatial context information. Furthermore, we propose a dynamic distillation fusion (DDF) module to enable dynamic signals aggregation and communication between hierarchical modules to further improve performance. Empirical results show that our DDistill-SR outperforms the baselines and achieves state-of-the-art results on most super-resolution domains with much fewer parameters and less computational overhead. We have released the code of DDistill-SR at <a class="link-external link-https" href="https://github.com/icandle/DDistill-SR" rel="external noopener nofollow">this https URL</a>.

An Efficient Feature Reuse Distillation Network for Lightweight Image Super-Resolution

Lightweight Image Super-Resolution Network Using 3D Convolutional Neural Networks

Epistemic-Uncertainty-Based Divide-and-Conquer Network for Single-Image Super-Resolution

Residual Feature Distillation Network for Lightweight Image Super-Resolution

Lightweight Multi-Attention Fusion Network for Image Super-Resolution

Feature Distillation Interaction Weighting Network for Lightweight Image Super-Resolution

CFDN: cross-scale feature distillation network for lightweight single image super-resolution

A Lightweight Image Super-Resolution Reconstruction Algorithm Based on the Residual Feature Distillation Mechanism

A Multi-Attention Feature Distillation Neural Network for Lightweight Single Image Super-Resolution

Lightweight Feature De-redundancy and Self-calibration Network for Efficient Image Super-resolution

Lightweight Single Image Super-resolution with Dense Connection Distillation Network

Deeply Recursive Low- and High-Frequency Fusing Networks for Single Image Super-Resolution

Efficient Image Super-Resolution with Feature Interaction Weighted Hybrid Network

Fast and Accurate Single Image Super-Resolution Via an Energy-Aware Improved Deep Residual Network.

Lightweight global-locally connected distillation network for single image super-resolution

Lightweight hierarchical residual feature fusion network for single-image super-resolution

DDistill-SR: Reparameterized Dynamic Distillation Network for Lightweight Image Super-Resolution

Fast and Accurate Single Image Super-Resolution via Information Distillation Network

Single-image Super-Resolution Via Selective Multi-Scale Network

Distilling with Residual Network for Single Image Super Resolution

Deep Fractal Residual Network for Fast and Accurate Single Image Super Resolution