Fast and accurate image upscaling with super-resolution forests

By Samuel Schulter, Christian Leistner, Horst Bischof. 2015

595 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): “The regression problem can also be solved with Gaussian process regression [26], trees [44] or Random Forests [45].”
Perceptual Losses for Real-Time Style Transfer and Super-Resolution (2016, neutral): “Other recent state-of-the-art methods include [39,40,41].”
Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network (2016, methodology): “For the interested reader, the results of other previous methods can be found in [31].”
Image Super-Resolution Using Deep Convolutional Networks (2014, methodology): “These methods either exploit internal similarities of the same image [5], [13], [16], [19], [47], or learn mapping functions from external low- and high-resolution exemplar pairs [2], [4], [6], [15], [23], [25], [37], [41], [42], [47], [48], [50], [51].”
Accurate Image Super-Resolution Using Very Deep Convolutional Networks (2015, methodology): “Ground Truth A+ [22] RFL [18] SelfEx [11] SRCNN [5] VDSR (Ours) (PSNR, SSIM) (27.”
Deeply-Recursive Convolutional Network for Image Super-Resolution (2015, methodology): “As some methods such as A+ [29] and RFL [23] do not predict image boundary, they require cropping pixels near borders.”
Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolution (2017, methodology): “We observe that methods using the bicubic upsampling for pre-processing generate results with noticeable artifacts [7, 17, 26, 30, 33].”
Learning a No-Reference Quality Metric for Single-Image Super-Resolution (2016, methodology): “Numerous single-image SR algorithms have been recently proposed for generic images that exploit priors based on edges [1], gradients [2, 3], neighboring interpolation [4, 5], regression [6], and patches [7, 8, 9, 10, 11, 12, 13, 14, 15].”
Fast and Accurate Single Image Super-Resolution via Information Distillation Network (2018, background): “This type of approaches usually focuses on how to learn a compact dictionary or manifold space to relate LR/HR patches, such as nearest neighbor [7], manifold embedding [2], random forest [20] and sparse representation [25, 26].”
Accelerating the Super-Resolution Convolutional Neural Network (2016, methodology): “First, we compare our method with four state-of-the-art learning-based SR algorithms that rely on external databases, namely the super-resolution forest (SRF) [7], SRCNN [1], SRCNN-Ex [2] and the sparse coding based network (SCN) [8].”
Super-resolution reconstruction of MR image with a novel residual learning network algorithm (2018, methodology): “Typical SR methods include the sparse representation-based SR algorithms (Rueda et al 2013, Wang et al 2014, Lu et al 2015, Zhang et al 2015, Jia et al 2016, 2017) and the regression learning-based algorithms (Alexander et al 2014, Schulter et al 2015, Jog et al 2016, Bahrami et al 2017).”
Balanced Two-Stage Residual Networks for Image Super-Resolution (2017, background): “Random forest can be built for SR without dictionary learning in the work of [29, 28].”
End-to-End Image Super-Resolution via Deep and Shallow Convolutional Networks (2016, methodology): “Based on the regression trees or forests algorithm, another line of image SR technique [12]–[14] is proposed, which builds on linear multivariate regression models using leaf nodes and locally linearizes the mapping from LR to HR”
Naive Bayes Super-Resolution Forest (2015, methodology): “[6] (SRCNN), the recently improved version of ANR [24] (A+) and the very recent ASRF with 15 trees of L = 16 layers [20] using the alternative training for better PSNR.”
Conditioned Regression Models for Non-blind Single Image Super-Resolution (2015, methodology): “In the original work of [24] this measure is defined as”
Image Restoration Using Convolutional Auto-encoders with Symmetric Skip Connections (2016, methodology): “We compare our network with SRCNN [28], NBSRF [53], CSCN [17], CSC [16], TSE [54] and ARFL+ [55] on three dataset: Set5, Set14 and BSD100.”