NTIRE 2018 Challenge on Single Image Super-Resolution: Methods and Results

By Radu Timofte, Shuhang Gu, Jiqing Wu, Luc Van Gool. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2018

354 citations recorded by the source corpus.

This paper reviews the 2nd NTIRE challenge on single image super-resolution (restoration of rich details in a low resolution image) with focus on proposed solutions and results. The challenge had 4 tracks. Track 1 employed the standard bicubic downscaling setup, while Tracks 2, 3 and 4 had realistic unknown downgrading operators simulating camera image acquisition pipeline. The operators were learnable through provided pairs of low and high resolution train images. The tracks had 145, 114, 101, and 113 registered participants, resp., and 31 teams competed in the final testing phase. They gauge the state-of-the-art in single image super-resolution.

Subfields: Atomic and Molecular Physics, and Optics, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/2c5726ce28186b5c8b275d1d2dc296d9db9528ca

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Atomic and Molecular Physics, and Optics

In Computer Vision and Pattern Recognition

Citations in the corpus

A Deep Journey into Super-resolution (2019, background): “, 1 and 2 distances or a combination of both [85].”
NTIRE 2019 Challenge on Video Deblurring and Super-Resolution: Dataset and Study (2019, methodology): “Still, as seen in the single image super-resolution literature [30, 8, 32, 1, 31, 2, 4] for further progress in video deblurring and SR, standardized benchmarks are essential to allow comparison of different proposed methods under the same conditions.”
Methods for image denoising using convolutional neural network: a review (2021, methodology): “They include: ImageNet large scale visual recognition challenge object detection (ILSVRCDET) [122], Places2 [123], BerkeleySegmentation Dataset (BSD) [124], waterloo explorationdatabase [125], EMNIST [126], COCO dataset [127], MIT-Adobe five k [128], ImageNet [129], BSD68 [130], Set 14 [131], Renoir”
Image Super-Resolution via Progressive Cascading Residual Network (2018, methodology): “(NTIRE2018 challenge [28] task1: ×8 scale)”
Efficient Transformer for Single Image Super-Resolution (2021, methodology): “It is worth noting that, training a Transformer usually needs a very large dataset, but SR datasets are usually very small (DIV2K [13] only has 1000 images).”
Frequency Separation for Real-World Super-Resolution (2019, background): “In recent years, the ﬁeld of image super-resolution has been dominated by CNNs, which achieve state-of-the-art performance [33, 34, 4].”
Transformer for Single Image Super-Resolution (2021, methodology): “In this work, we use DIV2K [35] as the training dataset.”
Contextual Residual Aggregation for Ultra High-Resolution Image Inpainting (2020, methodology): “SRGAN [37] is a learning-based method that can perform 4 × super-resolution and the ofﬁcial pre-trained model was trained on DIV2K.”
The 2018 PIRM Challenge on Perceptual Image Super-Resolution (2018, methodology): “, by PSNR, SSIM) [19, 11, 39, 36, 38] and in terms of visual quality (as rated by human observers) [18, 24, 31, 42, 44].”
Blind Super-Resolution Kernel Estimation using an Internal-GAN (2019, methodology): “Few Deep Blind-SR methods [34, 36] were recently presented in the NTIRE’2018 SR challenge [31].”
AI Benchmark: Running Deep Neural Networks on Android Smartphones (2018, background): “onnected to tasks meant for mobile devices. One notable group of such tasks is concerned with computer vision problems like image classiﬁcation [1,2,3], image enhancement [4,5,6] and super-resolution [7,8,9], optical character recognition [10], object tracking [11,12], visual scene understanding [13”
2018 PIRM Challenge on Perceptual Image Super-resolution (2018, methodology): “, NTIRE [38]) focus mostly on quantifying reconstruction accuracy, using e.”
Unsupervised Image Super-Resolution with an Indirect Supervised Path (2019, methodology): “Datasets DIV2K [1, 28], which contains 1,000 images with different scenes and is splitted to {800, 100, 100} for training, validation and testing.”
SRDiff: Single Image Super-Resolution with Diffusion Probabilistic Models (2021, neutral): “For general SR, we use the DIV2K [26] and Flickr2K [26] data-sets.”