Single Image Super-Resolution via Cascaded Multi-Scale Cross Network

The deep convolutional neural networks have achieved significant improvements in accuracy and speed for single image super-resolution. However, as the depth of network grows, the information flow is weakened and the training becomes harder and harder. On the other hand, most of the models adopt a single-stream structure with which integrating complementary contextual information under different receptive fields is difficult. To improve information flow and to capture sufficient knowledge for reconstructing the high-frequency details, we propose a cascaded multi-scale cross network (CMSC) in which a sequence of subnetworks is cascaded to infer high resolution features in a coarse-to-fine manner. In each cascaded subnetwork, we stack multiple multi-scale cross (MSC) modules to fuse complementary multi-scale information in an efficient way as well as to improve information flow across the layers. Meanwhile, by introducing residual-features learning in each stage, the relative information between high-resolution and low-resolution features is fully utilized to further boost reconstruction performance. We train the proposed network with cascaded-supervision and then assemble the intermediate predictions of the cascade to achieve high quality image reconstruction. Extensive quantitative and qualitative evaluations on benchmark datasets illustrate the superiority of our proposed method over state-of-the-art super-resolution methods.

• Publication year

  2018

• Venue

  arXiv.org

• Publication date

  2018-02-24

• Fields of study

  Computer Science, Engineering

• Identifiers
  arXiv 1802.08808
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• A joint residual network with paired ReLUs activation for image super-resolution

  2018cited by this paper
• MemNet: A Persistent Memory Network for Image Restoration

  2017influential reference
• Image Super Resolution Based on Fusing Multiple Convolution Neural Networks

  2017cited by this paper
• Image Super-Resolution Using Dense Skip Connections

  2017cited by this paper
• NTIRE 2017 Challenge on Single Image Super-Resolution: Methods and Results

  2017cited by this paper
• DiracNets: Training Very Deep Neural Networks Without Skip-Connections

  2017cited by this paper
• Image Super-Resolution via Deep Recursive Residual Network

  2017influential reference
• Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolution

  2017influential reference
• Fast and Accurate Image Super Resolution by Deep CNN with Skip Connection and Network in Network

  2017cited by this paper
• Enhanced Deep Residual Networks for Single Image Super-Resolution

  2017cited by this paper
• Image Restoration Using Very Deep Convolutional Encoder-Decoder Networks with Symmetric Skip Connections

  2016cited by this paper
• Accelerating the Super-Resolution Convolutional Neural Network

  2016cited by this paper
• Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning

  2016cited by this paper
• End-to-End Image Super-Resolution via Deep and Shallow Convolutional Networks

  2016cited by this paper
• Identity Mappings in Deep Residual Networks

  2016influential reference
• Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network

  2016cited by this paper
• Densely Connected Convolutional Networks

  2016cited by this paper
• SERF: A Simple, Effective, Robust, and Fast Image Super-Resolver From Cascaded Linear Regression

  2016cited by this paper
• Deep Convolutional Neural Networks with Merge-and-Run Mappings

  2016influential reference
• Single-Image Super-Resolution Using Active-Sampling Gaussian Process Regression

  2016cited by this paper
• Deeply-Recursive Convolutional Network for Image Super-Resolution

  2015influential reference
• Rethinking the Inception Architecture for Computer Vision

  2015cited by this paper
• Single image super-resolution from transformed self-exemplars

  2015cited by this paper
• Deep Residual Learning for Image Recognition

  2015cited by this paper
• Seven Ways to Improve Example-Based Single Image Super Resolution

  2015cited by this paper
• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

  2015cited by this paper
• Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification

  2015cited by this paper
• Training Very Deep Networks

  2015cited by this paper
• Accurate Image Super-Resolution Using Very Deep Convolutional Networks

  2015influential reference
• Deep Networks for Image Super-Resolution with Sparse Prior

  2015cited by this paper
• Fast and accurate image upscaling with super-resolution forests

  2015cited by this paper
• Deep Network Cascade for Image Super-resolution

  2014cited by this paper
• Image Super-Resolution Using Deep Convolutional Networks

  2014influential reference
• A+: Adjusted Anchored Neighborhood Regression for Fast Super-Resolution

  2014cited by this paper
• Going deeper with convolutions

  2014cited by this paper
• Beta Process Joint Dictionary Learning for Coupled Feature Spaces with Application to Single Image Super-Resolution

  2013cited by this paper
• Rectifier Nonlinearities Improve Neural Network Acoustic Models

  2013influential reference
• Image Super-Resolution With Sparse Neighbor Embedding

  2012cited by this paper
• Low-Complexity Single-Image Super-Resolution based on Nonnegative Neighbor Embedding

  2012influential reference
• A Bayesian Nonparametric Approach to Image Super-Resolution

  2012cited by this paper
• Contour Detection and Hierarchical Image Segmentation

  2011cited by this paper
• On Single Image Scale-Up Using Sparse-Representations

  2010cited by this paper
• Image Super-Resolution Via Sparse Representation

  2010influential reference
• An information fidelity criterion for image quality assessment using natural scene statistics

  2005influential reference
• Image quality assessment: from error visibility to structural similarity

  2004cited by this paper
• Super-resolution through neighbor embedding

  2004cited by this paper
• Learning Low-Level Vision

  1999cited by this paper

• Wind speed super-resolution and validation: from ERA5 to CERRA via diffusion models

  2024cites this paper
• DAW-GAN: a generative adversarial network based on the dynamic adaptive weight for image super-resolution

  2024cites this paper
• Multi-Scale Visual Perception Based Progressive Feature Interaction Network for Stereo Image Super-Resolution

  2024cites this paper
• Compatibility Review for Object Detection Enhancement through Super-Resolution

  2024influential citation
• Single image super-resolution based on trainable feature matching attention network

  2024cites this paper
• Improving Energy Consumption and Performance of IoT Based Camera Images Using Super-Resolution CNNs and Coarse-to-Fine Approach

  2024cites this paper
• Multiscale denoising generative adversarial network for speckle reduction in optical coherence tomography images

  2023cites this paper
• Fusing Attention Network Based on Dilated Convolution for Superresolution

  2023cites this paper
• A Multi-Scale Recursive Attention Feature Fusion Network for Image Super-Resolution Reconstruction Algorithm

  2023cites this paper
• A Review of Single Image Super Resolution Techniques using Convolutional Neural Networks

  2023cites this paper
• Dense channel splitting network for MR image super-resolution.

  2022cites this paper
• Study of Deep Learning-based models for Single Image Super-Resolution

  2022cites this paper
• Synthetic aperture radar image despeckling using convolutional neural networks in wavelet domain

  2022cites this paper
• Lightweight frequency-based attention network for image super-resolution

  2022cites this paper
• Single image super-resolution based on progressive fusion of orientation-aware features

  2022cites this paper
• Transpose convolution based model for super-resolution image reconstruction

  2022cites this paper
• A novel image super-resolution algorithm based on multi-scale dense recursive fusion network

  2022cites this paper
• Multiscale Recursive Feedback Network for Image Super-Resolution

  2022cites this paper
• Image Super-Resolution Based on Generalized Residual Network

  2021cites this paper
• Superresolution Imaging With a Deep Multipath Network for the Reconstruction of Satellite Cloud Images

  2021cites this paper
• Image super-resolution network based on a multi-branch attention mechanism

  2021cites this paper
• Progressive low/high-resolution Space Attention Fusion Network for Single Image Super-Resolution

  2021cites this paper
• A lightweight multi-scale feature integration network for real-time single image super-resolution

  2021cites this paper
• ESKN: Enhanced selective kernel network for single image super-resolution

  2021cites this paper
• Multi-scale attention network for image super-resolution

  2021cites this paper
• Lightweight feature separation, fusion and optimization networks for accurate image super-resolution

  2021cites this paper
• Lightweight multi-scale residual networks with attention for image super-resolution

  2020cites this paper
• VolumeNet: A Lightweight Parallel Network for Super-Resolution of MR and CT Volumetric Data

  2020cites this paper
• Lightweight image super-resolution with enhanced CNN

  2020cites this paper
• Cascaded dual-scale crossover network for hyperspectral image classification

  2020cites this paper
• Super-resolution using multi-channel merged convolutional network

  2020cites this paper
• Pyramid With Super Resolution for In-the-Wild Facial Expression Recognition

  2020cites this paper
• Improved dual-scale residual network for image super-resolution

  2020cites this paper
• Multi-Scale Residual Fusion Network for Super-Resolution Reconstruction of Single Image

  2020cites this paper
• Coarse-to-Fine CNN for Image Super-Resolution

  2020cites this paper
• VolumeNet: A Lightweight Parallel Network for Super-Resolution of Medical Volumetric Data

  2020cites this paper
• Image Super-Resolution Using Lightweight Multiscale Residual Dense Network

  2020cites this paper
• Single Image Super-Resolution via Cascaded Parallel Multisize Receptive Field

  2019cites this paper
• Multilevel and Multiscale Network for Single-Image Super-Resolution

  2019cites this paper
• Gibbs-ringing artifact suppression with knowledge transfer from natural images to MR images

  2019influential citation
• A Deep Journey into Super-resolution

  2019influential citation
• Pulmonary nodule image super-resolution using multi-scale deep residual channel attention network with joint optimization

  2019cites this paper
• Single Image Super-Resolution Based on Global Dense Feature Fusion Convolutional Network

  2019influential citation
• Deep Neural Network for Fast and Accurate Single Image Super-Resolution via Channel-Attention-based Fusion of Orientation-aware Features

  2019cites this paper
• Depth-Controllable Very Deep Super-Resolution Network

  2019cites this paper
• Densely Residual Laplacian Super-Resolution

  2019cites this paper
• Multi-Grained Attention Networks for Single Image Super-Resolution

  2019cites this paper
• Deep learning based super resolution of aerial and satellite imagery

  2019cites this paper
• Generative adversarial networks based super resolution of satellite aircraft imagery

  2019cites this paper
• DEEP LEARNING FOR IMAGE RESTORATION AND ROBOTIC VISION

  2019cites this paper
• HiCNN: a very deep convolutional neural network to better enhance the resolution of Hi-C data

  2019cites this paper
• Single MR Image Super-Resolution via Channel Splitting and Serial Fusion Network

  2019influential citation
• Super-Resolution Reconstruction Algorithm for Cross-Module Based on Depth-Wise Separable Convolution

  2018cites this paper
• Channel Splitting Network for Single MR Image Super-Resolution

  2018cites this paper