Fast deep parallel residual network for accurate super resolution image processing

A Combined Model of Diffusion Model and Enhanced Residual Network for Super-Resolution Reconstruction of Turbulent Flows

“…p(y T ) = N (y T | 0, I) (10) If the noise variances in the forward process steps are set as small as possible, i.e., α 1:T ≈ 1, the optimal reverse process p θ 1 (y t−1 | y t , x) will approximate a Gaussian distribution [31].…”

“…Inspired by various research advances in the super-resolution deep learning of images, such as the convolutional neural network-based (CNN-based) methods [6][7][8][9][10], the generative adversarial network-based (GAN-based) methods [11][12][13], the transformer-based methods [14,15], and the diffusion-based methods [16][17][18], several neural network models have been proposed to reconstruct a high-resolution (HR) flow field. We categorize the recent super-resolution (SR) methods for flow fields into three categories:…”

Self-supervised cycle-consistent learning for scale-arbitrary real-world single image super-resolution

“…Existing SR methods on synthetic images can be roughly grouped into two categories, i.e., reconstruction-based methods (Chang et al, 2019;Chen et al, 2017;Dong et al, 2013Dong et al, , 2011Jiang et al, 2017;Li et al, 2018;Zhang et al, 2012) and learning-based methods (Chang et al, 2020;Dai et al, 2019;Dong et al, 2015;Guo et al, 2020;Huang et al, 2017Huang et al, , 2021Ledig et al, 2017;Li et al, 2020;Lim et al, 2017;Perez-Pellitero et al, 2016;Schulter et al, 2015;Sha et al, 2019;Sharma and Kumar, 2021;Timofte et al, 2013Timofte et al, , 2014Wang et al, 2022;Yang et al, 2010;Zeyde et al, 2010;Zhang et al, 2019aZhang et al, , 2021bZhang et al, , 2018bZhao et al, 2019;Zhi-Song and Siu, 2018;Zhou et al, 2021). By modeling the degradation process of LR images and the properties of HR natural images, reconstruction-based methods covert the SISR task into a minimization problem under the MAP estimation framework.…”

“…Learning-based methods generally obtain the LR-to-HR mapping from training images. A variety of models have been used to learn the mapping, including sparse coding (Li et al, 2020;Yang et al, 2010;Zeyde et al, 2010), neighborhood regression (Perez-Pellitero et al, 2016;Timofte et al, 2013Timofte et al, , 2014Zhang et al, 2019a), random forests (Huang et al, 2017;Schulter et al, 2015;Zhi-Song and Siu, 2018), and deep CNNs (Chang et al, 2020;Dai et al, 2019;Dong et al, 2015;Guo et al, 2020;Huang et al, 2021;Ledig et al, 2017;Lim et al, 2017;Sha et al, 2019;Sharma and Kumar, 2021;Wang et al, 2022;Zhang et al, 2021bZhang et al, , 2018bZhao et al, 2019;Zhou et al, 2021). Among them, deep CNNs are the current mainstream models for their impressive reconstruction accuracy and high efficiency on the GPU platform.…”

“…In addition to modeling natural image properties, priors can also be learned from datasets. Sparse coding (Li et al, 2020;Yang et al, 2010;Zeyde et al, 2010), neighborhood regression (Perez-Pellitero et al, 2016;Timofte et al, 2013Timofte et al, , 2014Zhang et al, 2019a), random forests (Huang et al, 2017;Schulter et al, 2015;Zhi-Song and Siu, 2018), and deep convolutional neural networks (CNNs) (Chang et al, 2020;Dai et al, 2019;Dong et al, 2015;Guo et al, 2020;Huang et al, 2021;Ledig et al, 2017;Lim et al, 2017;Sha et al, 2019;Sharma and Kumar, 2021;Wang et al, 2022;Zhang et al, 2021bZhang et al, , 2018bZhao et al, 2019;Zhou et al, 2021) are adopted in previous studies to learn the mapping from LR images to their HR counterparts. In particular, as in many other fields (Barzekar and Yu, 2022;Chala et al, 2021;Dai et al, 2022;Garg et al, 2022;Jiang et al, 2022;Liu et al, 2021;Tan et al, 2021;Zangeneh et al, 2020), deep CNN-based models have uplifted the SR performance to a new height.…”