Coupled Deep Autoencoder for Single Image Super-Resolution

  title={Coupled Deep Autoencoder for Single Image Super-Resolution},
  author={Kun Zeng and Jun Yu and Ruxin Wang and Cuihua Li and Dacheng Tao},
  journal={IEEE Transactions on Cybernetics},
Sparse coding has been widely applied to learning-based single image super-resolution (SR) and has obtained promising performance by jointly learning effective representations for low-resolution (LR) and high-resolution (HR) image patch pairs. However, the resulting HR images often suffer from ringing, jaggy, and blurring artifacts due to the strong yet ad hoc assumptions that the LR image patch representation is equal to, is linear with, lies on a manifold similar to, or has the same support… 
Deep learning based image super-resolution with coupled backpropagation
Preliminary experiments show that SR-CBP can produce better results than state of the art single image superresolution methods based on sparse representations, and the memory and storage requirements of CAN are lesser than existing deep learning based SR methods.
Coupled deep auto-encoder with image edge information for image super-resolution
Expensive experimentation results show that the proposed algorithm generates higher-quality super-resolution image compared to other state-of-the-art methods.
Structure-Preserving Image Super-Resolution via Contextualized Multitask Learning
This paper develops a contextualized multitask learning framework that first extracts convolutional features from the input LR image and applies one deconvolutional module to interpolate the LR feature maps in a content-adaptive way and demonstrates the effectiveness of the method on achieving both higher restoration quality and computational efficiency compared with several state-of-the-art SR approaches.
Single Image Super-Resolution With Learning Iteratively Non-Linear Mapping Between Low- and High-Resolution Sparse Representations
This work first learns LR dictionary, and then obtains HR dictionary and a nonlinear mapping between LR/HR sparse domains by training them iteratively, which can reveals accurately the intrinsic non-linear relationship between LR and HR sparse domains.
Single Image Internal Distribution Measurement Using Non-Local Variational Autoencoder
NLVAE is introduced as a self-supervised strategy that re-constructs high-resolution images using disentangled information from the non-local neighbourhood that outperforms a number of baseline and state-of-the-art methods through extensive qualitative and quantitative evaluations.
Deep learning based frameworks for image super-resolution and noise-resilient super-resolution
Experimental results show that proposed noise resilient super-resolution framework outperforms the conventional and state-of-the-art approaches in terms of PSNR and SSIM metrics.
Coupled Autoencoder Network with Joint Regularizations for image super-resolution
A sparse deep autoencoder network with joint regularizations for image super-resolution is built due to the local smoothness and the redundancy of an image, and the joint variation regularizations are unified with the coupled autoen coder network (CAN).
Single image super-resolution reconstruction via combination mapping with sparse coding
A new learning-based method of single-image super-resolution reconstruction from single low-resolution (LR) image, Inspired by simple mapping functions method, where jagged and blurred artifacts are depressed effectively and high reconstruction quality is acquired with less exemplar images.
Single-image super-resolution via joint statistic models-guided deep auto-encoder network
A unified model which combines the deep model with the image restoration model for single-image SR is proposed which can not only reconstruct the SR image, but also keep the distinct fine structures for the low-resolution image.


A Unified Learning Framework for Single Image Super-Resolution
This paper proposes a new SR framework that seamlessly integrates learning- and reconstruction-based methods for single image SR to avoid unexpected artifacts introduced by learning-based SR and restore the missing high-frequency details smoothed by reconstruction- based SR.
Coupled Dictionary Training for Image Super-Resolution
This paper demonstrates that the coupled dictionary learning method can outperform the existing joint dictionary training method both quantitatively and qualitatively and speed up the algorithm approximately 10 times by learning a neural network model for fast sparse inference and selectively processing only those visually salient regions.
Multi-scale dictionary for single image super-resolution
This work introduces a multi-scale dictionary to a novel SR method that simultaneously integrates local and non-local priors and demonstrates that the proposed method can produce high quality SR recovery both quantitatively and perceptually.
Single-Image Super-Resolution Using Sparse Regression and Natural Image Prior
  • K. Kim, Younghee Kwon
  • Computer Science, Mathematics
    IEEE Transactions on Pattern Analysis and Machine Intelligence
  • 2010
Compared with existing algorithms, KRR leads to a better generalization than simply storing the examples as has been done in existing example-based algorithms and results in much less noisy images.
Learning a Deep Convolutional Network for Image Super-Resolution
This work proposes a deep learning method for single image super-resolution (SR) that directly learns an end-to-end mapping between the low/high-resolution images and shows that traditional sparse-coding-based SR methods can also be viewed as a deep convolutional network.
Image super-resolution as sparse representation of raw image patches
It is shown that a small set of randomly chosen raw patches from training images of similar statistical nature to the input image generally serve as a good dictionary, in the sense that the computed representation is sparse and the recovered high-resolution image is competitive or even superior in quality to images produced by other SR methods.
A Shrinkage Learning Approach for Single Image Super-Resolution with Overcomplete Representations
A novel approach for online shrinkage functions learning in single image super-resolution that leverages the classical Wavelet Shrinkage denoising technique and demonstrates superior performance compared to state-of-the-art results.
Semi-coupled dictionary learning with applications to image super-resolution and photo-sketch synthesis
The proposed semi-coupled dictionary learning (SCDL) model is applied to image super-resolution and photo-sketch synthesis, and the experimental results validated its generality and effectiveness in cross-style image synthesis.
Super resolution using edge prior and single image detail synthesis
This paper proposes an approach to extend edge-directed super-resolution to include detail from an image/texture example provided by the user (e.g., from the Internet), and can achieve quality results at very large magnification, which is often problematic for both edge- directed and learning-based approaches.
Fast Image Super-Resolution Based on In-Place Example Regression
We propose a fast regression model for practical single image super-resolution based on in-place examples, by leveraging two fundamental super-resolution approaches- learning from an external