NTIRE 2017 Challenge on Single Image Super-Resolution: Dataset and Study

Enhanced Super-Resolution Generative Adversarial Network (ESRGAN) is a perceptual-driven approach for single image super resolution that is able to produce photorealistic images. Despite the visual quality of these generated images, there is still room for improvement. In this fashion, the model is extended to further improve the perceptual quality of the images. We have designed a novel block to replace the one used by the original ESRGAN. Moreover, we introduce noise inputs to the generator network in order to exploit stochastic variation. The resulting images present more realistic textures.

show abstract

“…The used training set is DIV2K [20]. It is a dataset of 2K resolution images adequate for the task of SR.…”

Section: Datamentioning

confidence: 99%

ESRGAN+ : Further Improving Enhanced Super-Resolution Generative Adversarial Network

Rakotonirina

Rasoanaivo

2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

151

View full text Add to dashboard Cite

show abstract

“…For fair comparisons, we re-train the models of SRResNet , ED-DSRN , and PFFNet on our training set 1 . Other super resolution methods are trained on the DIV2K dataset (Agustsson and Timofte, 2017) and deep learning-based dehazing methods are trained on the RESIDE dataset (Li et al, 2017b). The quantitative evaluations in Table 2 show that the proposed GFN model performs well in terms of PSNR and SSIM with shorter inference time.…”

Section: Super Resolving Hazy Imagementioning

confidence: 99%

Gated Fusion Network for Degraded Image Super Resolution

Zhang

Dong

Hu³

et al. 2020

Int J Comput Vis

View full text Add to dashboard Cite

Single image super resolution aims to enhance image quality with respect to spatial content, which is a fundamental task in computer vision. In this work, we address the task of single frame super resolution with the presence of image degradation, e.g., blur, haze, or rain streaks. Due to the limitations of frame capturing and formation processes, image degradation is inevitable, and the artifacts would be exacerbated by super resolution methods. To address this problem, we propose a dual-branch convolutional neural network to extract base features and recovered features separately. The base features contain local and global information of the input image. On the other hand, the recovered features focus on the degraded regions and are used to remove the degradation. Those features are then fused through a recursive gate module to obtain sharp features for super resolution. By decomposing the feature extraction step into two task-independent streams, the dual-branch model can facilitate the training process by avoiding learning the mixed degradation all-in-one and thus enhance the final high-resolution prediction results. We evaluate the proposed method in three degradation scenarios. Experiments on these scenarios demonstrate that the proposed method performs more efficiently and favorably against the state-of-the-art approaches on benchmark datasets.

show abstract

“…This part introduces an ablation experiment we have conducted. In the experiment, a lightweight EDSR [12] network [19], which consists of 800 training images and 100 validation images. As shown in Fig.…”

Section: Ablation Experimentsmentioning

confidence: 99%

Pixel-Level Self-Paced Learning For Super-Resolution

Lin

Wang

et al. 2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Recently, lots of deep networks are proposed to improve the quality of predicted super-resolution (SR) images, due to its widespread use in several image-based fields. However, with these networks being constructed deeper and deeper, they also cost much longer time for training, which may guide the learners to local optimization. To tackle this problem, this paper designs a training strategy named Pixel-level Self-Paced Learning (PSPL) to accelerate the convergence velocity of SISR models. PSPL imitating self-paced learning gives each pixel in the predicted SR image and its corresponding pixel in ground truth an attention weight, to guide the model to a better region in parameter space. Extensive experiments proved that PSPL could speed up the training of SISR models, and prompt several existing models to obtain new better results. Furthermore, the source code is available at https://github.com/Elin24/PSPL Index Termssuper-resolution, training strategy, selfpaced learning * Qi Wang is the corresponding author.

show abstract

NTIRE 2017 Challenge on Single Image Super-Resolution: Dataset and Study

Cited by 3,034 publications

References 43 publications

ESRGAN+ : Further Improving Enhanced Super-Resolution Generative Adversarial Network

ESRGAN+ : Further Improving Enhanced Super-Resolution Generative Adversarial Network

Gated Fusion Network for Degraded Image Super Resolution

Pixel-Level Self-Paced Learning For Super-Resolution

Contact Info

Product

Resources

About