Image upsampling via imposed edge statistics

By Raanan Fattal. ACM Transactions on Graphics, 2007

535 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/5d1c200b58b9b33fb8b87789339d39887077a7c2

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

Citations in the corpus

Single-Image Super-Resolution: A Benchmark (2014, background): “While domain-specific SISR algorithms focus on specific classes of images such as faces [35,42], scenes [33], and graphics artwork [18], generic SISR algorithms [10,38,8,3,27,30,11,46,9,34,32,12,5,39,44,43] are developed for all kinds of images where the priors are typically based on primitive image”
Image super-resolution using gradient profile prior (2008, methodology): “Fattal [7] also proposed an edge statistics for image upsampling.”
Exploiting the Potential of Standard Convolutional Autoencoders for Image Restoration by Evolutionary Search (2018, background): “In order to resolve or mitigate its ill-posedness, researchers have incorporated image priors such as edge statistics (Fattal, 2007), total variation (Perrone & Favaro, 2014), and sparse representation (Aharon et al., 2006; Yang et al., 2010), which are built on intuition or statistics of natural…”
Image and video upscaling from local self-examples (2011, methodology): “Despite being an examples-based method, our running times are also lower than the times reported for methods that use analytical models [Fattal 2007; Shan et al. 2008; Sun et al. 2008] on the same or stronger machines.”
CLEARER: Multi-Scale Neural Architecture Search for Image Restoration (2020, methodology): “To recover the clean image from the observed noisy image, a huge number of image restoration methods [30, 1, 5, 10] have been proposed based on different handcrafted priors.”
Deep Networks for Image Super-Resolution with Sparse Prior (2015, background): “Analytical priors, such as bicubic interpolation, work well for smooth regions; while image models based on statistics of edges [11] and gradients [17, 1] can recover sharper structures.”
Unsupervised Image Super-Resolution Using Cycle-in-Cycle Generative Adversarial Networks (2018, background): “Early approaches either rely on natural image statistics [33] [13] or pre-defined models [10] [5] [26].”
Deep Plug-And-Play Super-Resolution for Arbitrary Blur Kernels (2019, unknown): “, RCAN [67], SRMD [63] and ZSSR [51]) on classical image “chip” [22], noisy image “frog” [33] and blurry image “colour” [43].”
Recovering Realistic Texture in Image Super-Resolution by Deep Spatial Feature Transform (2018, methodology): “Image priors such as edge features [13, 41], statistics [24, 1] and internal patch recurrence [16] are employed to improve performance.”
Super-resolution from a single image (2009, methodology): “Our approach attempts to recover at each pixel its best possible resolution increase based on its patch redundancy within and across scales.”
A Deep Convolutional Neural Network with Selection Units for Super-Resolution (2017, methodology): “Fattal [14] utilized edge statistics of LR images to reconstruct their sharper HR counterparts.”
Structure-Preserving Super Resolution With Gradient Guidance (2020, methodology): “For SR methods, Fattal [11] proposes a method based on edge statistics of image gradients by learning the prior dependency of different resolutions.”
Single-Image Super-Resolution Using Sparse Regression and Natural Image Prior (2010, unknown): “As a regularized solution, 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.”
Multi-scale dictionary for single image super-resolution (2012, methodology): “For example, Fattal [6] enforced an edge statistics prior on the reconstructed HR image to generate sharper edges and suppress noticeable artifacts.”
Pixel Recursive Super Resolution (2017, background): “Enhancing linear methods with rich image priors such as sparsity [2] or Gaussian mixtures [51] have substantially improved the quality of the methods; likewise, leveraging low-level image statistics such as edge gradients improves predictions [47, 41, 8, 16, 39, 22].”
Unfolding the Alternating Optimization for Blind Super Resolution (2020, methodology): “As shown in Figure 6 (a) and (c), the average PSNR results on Set5 and Set14 ﬁrstly increase rapidly and then gradually converge.”