SimUSR: A Simple but Strong Baseline for Unsupervised Image Super-resolution

@article{Ahn2020SimUSRAS,
  title={SimUSR: A Simple but Strong Baseline for Unsupervised Image Super-resolution},
  author={Namhyuk Ahn and Jaejun Yoo and Kyung-ah Sohn},
  journal={2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)},
  year={2020},
  pages={1953-1961}
}
  • Namhyuk Ahn, Jaejun Yoo, Kyung-ah Sohn
  • Published 23 April 2020
  • Computer Science, Environmental Science
  • 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
In this paper, we tackle a fully unsupervised super-resolution problem, i.e., neither paired images nor ground truth HR images. We assume that low resolution (LR) images are relatively easy to collect compared to high resolution (HR) images. By allowing multiple LR images, we build a set of pseudo pairs by denoising and downsampling LR images and cast the original unsupervised problem into a supervised learning problem but in one level lower. Though this line of study is easy to think of and… 

Figures and Tables from this paper

Flexible Style Image Super-Resolution using Conditional Objective

TLDR
This work optimize an SR model with a conditional objective during training, where the objective is a weighted sum of multiple perceptual losses at different feature levels, and presents an architecture appropriate for this training scheme, which is the Residual-in-Residual Dense Block equipped with spatial feature transformation layers.

NTIRE 2020 Challenge on Real-World Image Super-Resolution: Methods and Results

TLDR
The NTIRE 2020 challenge addresses the real world setting, where paired true high and low-resolution images are unavailable, and the ultimate goal is to achieve the best perceptual quality, evaluated using a human study.

Progressive Splitting and Upscaling Structure for Super-Resolution

TLDR
A novel progressive splitting and upscaling structure is proposed, termed PSUS, which generates decoupled feature maps for upscaled layer to get better SR image and can achieve considerable improvement on image quality with fewer parameters and lower computational cost.

Learning Multiple Probabilistic Degradation Generators for Unsupervised Real World Image Super Resolution

TLDR
This paper shows that it can improve the performance of USR models by relaxing the assumption and proposing to train the probabilistic degradation generator, which can be viewed as a deep hierarchical latent variable model and is more suitable for modeling the complex conditional distribution.

EDKE: Encoder-Decoder based Kernel Estimation for Blind Image Super-resolution

TLDR
A novel encoder-decoder based kernel estimation (EDKE) framework by considering the anisotropy of kernel is proposed, which shows the superiority of the method over other state-of-the-art methods in kernel estimation.

CVPRW 2020 TOC

  • 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
  • 2020

References

SHOWING 1-10 OF 34 REFERENCES

Unsupervised Learning for Real-World Super-Resolution

TLDR
This work learns to invert the effects of bicubic downsampling in order to restore the natural image characteristics present in the data, and can be trained with direct pixel-wise supervision in the high resolution domain, while robustly generalizing to real input.

"Zero-Shot" Super-Resolution Using Deep Internal Learning

TLDR
This paper exploits the internal recurrence of information inside a single image, and train a small image-specific CNN at test time, on examples extracted solely from the input image itself, which is the first unsupervised CNN-based SR method.

Unsupervised Image Super-Resolution Using Cycle-in-Cycle Generative Adversarial Networks

TLDR
This work proposes a Cycle-in-Cycle network structure with generative adversarial networks (GAN) as the basic component to tackle the single image super-resolution problem in a more general case that the low-/high-resolution pairs and the down-sampling process are unavailable.

Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Analysis and a New Strategy

TLDR
CutBlur is proposed that cuts a low-resolution patch and pastes it to the corresponding high-resolution image region and vice versa and consistently and significantly improves the performance across various scenarios, especially when the model size is big and the data is collected under real-world environments.

Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network

TLDR
SRGAN, a generative adversarial network (GAN) for image super-resolution (SR), is presented, to its knowledge, the first framework capable of inferring photo-realistic natural images for 4x upscaling factors and a perceptual loss function which consists of an adversarial loss and a content loss.

Single image super-resolution from transformed self-exemplars

TLDR
This paper expands the internal patch search space by allowing geometric variations, and proposes a compositional model to simultaneously handle both types of transformations to accommodate local shape variations.

Meta-Transfer Learning for Zero-Shot Super-Resolution

TLDR
Meta-Transfer Learning for Zero-Shot Super-Resolution (MZSR) is presented, which leverages ZSSR and can exploit both external and internal information, where one single gradient update can yield quite considerable results.

Accurate Image Super-Resolution Using Very Deep Convolutional Networks

TLDR
This work presents a highly accurate single-image superresolution (SR) method using a very deep convolutional network inspired by VGG-net used for ImageNet classification and uses extremely high learning rates enabled by adjustable gradient clipping.

AIM 2019 Challenge on Real-World Image Super-Resolution: Methods and Results

TLDR
The AIM 2019 challenge on real world super-resolution addresses the real world setting, where paired true high and low-resolution images are unavailable, and aims to advance the state-of-the-art and provide a standard benchmark for this newly emerging task.

NTIRE 2020 Challenge on Real-World Image Super-Resolution: Methods and Results

TLDR
The NTIRE 2020 challenge addresses the real world setting, where paired true high and low-resolution images are unavailable, and the ultimate goal is to achieve the best perceptual quality, evaluated using a human study.