Single image super-resolution from transformed self-exemplars

By Jia-Bin Huang, Abhishek Singh, Narendra Ahuja. 2015

3,414 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network (2016, methodology): “We validate the proposed approaches using images from publicly available benchmark datasets and compare our performance against state-of-the-art works including SRCNN [8], SelfExSR [24] and DRCN [27].”
Perceptual Losses for Real-Time Style Transfer and Super-Resolution (2016, methodology): “We evaluate all models on the standard Set5 [65], Set14 [66], and BSD100 [46] datasets.”
VEGA: Towards an End-to-End Configurable AutoML Pipeline (2020, methodology): “We report the Peak signal-to-noise ratio (PSNR) on Urban100 dataset (Huang et al., 2015).”
AutoGAN-Distiller: Searching to Compress Generative Adversarial Networks (2020, neutral): “…end for 12: # Second Step: Search 13: for t ← 1 to T 2 do 14: Get a batch of data X 1 from χ 1 15: Get a batch of data X from χ 27: w ) 29: end for SR benchmarks, including Set5 (Bevilacqua et al., 2012), Set14 (Zeyde et al., 2010), BSD100 (Martin et al., 2001) and Urban100 (Huang et al., 2015).”
Hierarchical Neural Architecture Search for Single Image Super-Resolution (2020, methodology): “To show the effectiveness of out method, we conduct comparisons on ﬁve benchmark datasets, including Set5 [36], Set14 [37], BSD100 [38], Urban100 [39], and Manga109 [40].”
Soft-to-Hard Vector Quantization for End-to-End Learning Compressible Representations (2017, methodology): “Urban100 [14] has 100 color images selected from Flickr with labels such as urban, city, architecture, and structure.”
The Perception-Distortion Tradeoff (2017, methodology): “The evaluated algorithms include: A+ [66], SRCNN [67], SelfEx [68], VDSR [69], Johnson et al.”
Image Super-Resolution Using Deep Convolutional Networks (2014, background): “Index Terms—Super-resolution, deep convolutional neural networks, sparse coding F”
EnhanceNet: Single Image Super-Resolution Through Automated Texture Synthesis (2016, background): “In recent years, popular approaches include exemplarbased models that either exploit recurrent patches of different scales within a single image [13, 17, 22, 56] or learn mappings between low and high resolution pairs of image patches in external databases [3, 5, 14, 27, 51, 58, 63].”
ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks (2018, methodology): “B channels and augment the training dataset with random horizontal ips and 90 degree rotations. We evaluate our models on widely used benchmark datasets { Set5 [42], Set14 [43], BSD100 [44], Urban100 [45], and the PIRM self-validation dataset that is provided in the PIRM-SR Challenge. 4.3 Qualitativ”
Enhanced Deep Residual Networks for Single Image Super-Resolution (2017, methodology): “We also compare the performance on four standard benchmark datasets: Set5 [2], Set14 [33], B100 [17], and Urban100 [10].”
Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolution (2017, methodology): “We carry out extensive experiments using 5 datasets: SET5 [2], SET14 [39], BSDS100 [1], URBAN100 [15] and MANGA109 [23].”
Image Super-Resolution Using Very Deep Residual Channel Attention Networks (2018, positive): “For Urban100 and Manga109, the PSNR gains of RCAN over EDSR are 0.49 dB and 0.55 dB. EDSR has much larger number of parameters (43 M) than ours (16 M), but our RCAN obtains much better performance.”
Pixel Recursive Super Resolution (2017, background): “Much work has been done on algorithms that search a database of patches and combine them to create plausible high frequency details in zoomed images [7, 13].”
Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising (2016, methodology): “We adopt the four testing datasets (i.e., Set5 and Set14, BSD100 and Urban100 [40]) used in [35].”
A Deep Journey into Super-resolution (2019, methodology): “In this section, we compare the state-of-the-art algorithms on publicly available benchmark datasets which include Set5 [88], Set14 [89], BSD100 [90], Urban100 [91], DIV2K [50] and Manga109 [92].”
Residual Non-local Attention Networks for Image Restoration (2019, background): “As shown in Table 6, our RNAN+ achieves the second best performance among benchmark datasets: Set5 (Bevilacqua et al., 2012), Set14 (Zeyde et al., 2010), B100 (Martin et al., 2001), Urban100 (Huang et al., 2015), and Manga109 (Matsui et al., 2017).”
Wavelet-SRNet: A Wavelet-Based CNN for Multi-scale Face Super Resolution (2017, background): “In general, image super-resolution methods can be divided into three types: interpolation-based, statisticsbased [26, 31, 32] and learning-based methods [3, 9, 24].”
Recurrent Inference Machines for Solving Inverse Problems (2017, methodology): “nly about 500:000 parameters. 6We reimplemented MATLABs bicubic interpolation kernel in order to apply a forward model (sub-sampling) in TensorFlow which agrees with the forward model in Huang et al. [27]. 7https://sites.google.com/site/jbhuang0604/publications/struct_sr 7 Table 3: Color image super”