Super-Resolution Without Explicit Subpixel Motion Estimation

By Hiroyuki Takeda, Peyman Milanfar, Matan Protter, Michael Elad. IEEE Transactions on Image Processing, 2009

334 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/1056e2f995f9f205dfbad6faf1d1e2994a82230b

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

Bidirectional Recurrent Convolutional Networks for Multi-Frame Super-Resolution (2015, methodology): “55 3DSKR [42] 4224 1211 957 255 1580 323 443 127 697 173 1580 418 Enhancer [1] - - - - - - - - - - - BRCN 4.”
Detail-Revealing Deep Video Super-Resolution (2017, methodology): “Most previous methods compensate inter-frame motion by estimating optical flow [2, 7, 19, 20, 23] or applying block-matching [28].”
EDVR: Video Restoration With Enhanced Deformable Convolutional Networks (2019, methodology): “Earlier studies [36, 4, 33, 19, 11] treat video restoration as a simple extenBicubic RCAN”
Static and space-time visual saliency detection by self-resemblance. (2009, methodology): “For a more in depth analysis of loc al steering kernels, we refer the interested reader to [34], [35].”
On Bayesian Adaptive Video Super Resolution (2014, methodology): “Inspired by the successful nonlocal means method for video denoising, Takeda et al. [36] avoided explicit subpixel motion estimation and used 3D kernel regression to exploit the spatiotemporal neighboring relationship for video upsampling.”
Ieee Transactions on Pattern Analysis and Machine Intelligence 1 on Bayesian Adaptive Video Super Resolution (methodology): “In-depth experiments demonstrate that our system outperforms the state-of-the-art super resolution systems [1], [31], [36] on challenging real-world sequences.”
Frame-Recurrent Video Super-Resolution (2018, background): “Classical video and multi-frame super-resolution methods are generally formulated as optimization problems that are computationally very expensive to solve [2, 11, 27, 38].”
Robust Super-resolution Volume Reconstruction from Slice Acquisitions: Application to Fetal Brain MRI (2010, background): “Also, a few groups have recently developed super-resolution reconstruction techniques that are robust to motion estimation errors [27], [28].”
Cardiac Image Super-Resolution with Global Correspondence Using Multi-Atlas PatchMatch (2013, methodology): “Takeda [6] proposed a distance estimation approach which requires no explicit motion estimation by inverting the position of the patch selection operator and the resampling operator.”
Super-resolution: a comprehensive survey (2014, background): “– [318], [426], [460], [464], [468] they are defined based on steering kernel regression which takes into account the correlation between the pixel positions”
Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network (2016, methodology): “For video experiments we use 1080p HD videos from the publicly available Xiph database(1), which has been used to report video SR results in previous methods [37, 23].”
Video Super-Resolution With Convolutional Neural Networks (2016, methodology): “An alternative method to the conventional motion estimation and image restoration scheme is presented in [27].”
BasicVSR: The Search for Essential Components in Video Super-Resolution and Beyond (2020, background): “Video super-resolution, by nature, involves a long and complex processing pipeline since it needs to aggregate information from not only the spatial dimension but also the temporal dimension.”
Real-Time Video Super-Resolution with Spatio-Temporal Networks and Motion Compensation (2016, methodology): “A natural choice has been to preprocess input frames by compensating interframe motion using displacement fields obtained from offthe-shelf optical flow algorithms [36].”