Ultra-Resolving Face Images by Discriminative Generative Networks

By Xin Yu, Fatih Murat Porikli. 2016

306 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

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

Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network (2016, methodology): “ruction. In Mathieu et al. [42] and Denton et al. [7] the authors tackled this problem by employing generative adversarial networks (GANs) [22] for the application of image generation. Yu and Porikli [66] augment pixel-wise MSE loss with a discriminator loss to train a network that super-resolves fa”
EnhanceNet: Single Image Super-Resolution Through Automated Texture Synthesis (2016, background): “…include dictionary-based methods [33, 40, 52, 59, 61, 64] that learn a sparse representation of image patches as a combination of dictionary atoms, as well as neural network-based approaches [4, 8, 9, 24, 25, 26, 47, 48, 62] which apply convolutional neural networks (CNNs) to the task of SISR.”
Pixel Recursive Super Resolution (2017, methodology): “; a nearest neighbor search baseline inspired by previous work on example-based super resolution [9], (2) ResNet L2; a deep neural network using Resnet blocks trained with MSE objective, and (3) GAN; a GAN based super resolution model implemented by [11] similar to [49].”
Deep Learning for Image Super-Resolution: A Survey (2019, methodology): “The UR-DGN [177] adopts a network similar to SRGAN [25] with adversarial learning.”
To learn image super-resolution, use a GAN to learn how to do image degradation first (2018, result): “The recent work of [22] is a GAN-based approach similar to the one proposed in [2].”
Fast and Accurate Image Super-Resolution with Deep Laplacian Pyramid Networks (2017, methodology): “been applied to several image reconstruction and synthesis problems, including image inpainting [36], face completion [37], and face super-resolution [38].”
Learning to Super-Resolve Blurry Face and Text Images (2017, background): “Because strong image priors can be learned, GANs have been applied to image enhancement tasks such as face hallucination [50] and super-resolution [22].”
KGSNet: Key-Point-Guided Super-Resolution Network for Pedestrian Detection in the Wild (2020, background): “However, as we all know, the solution to the MSE optimization problem usually lacks high-frequency details, which results in blurred images. b) Body structure loss: As mentioned in [73] and [74], only using pixelwise MSE loss will produce images with oversmooth texture.”
FSRNet: End-to-End Learning Face Super-Resolution with Facial Priors (2017, methodology): “We compare FSRNet with state-of-the-art SR methods, including generic SR methods like SRResNet [22], VDSR [17] and SRCNN [6]; and facial SR methods like GLN [35] and URDGN [41].”
SRFlow: Learning the Super-Resolution Space with Normalizing Flow (2020, methodology): “This was addressed in UR-DGN [51], SRGAN [24] and more recent approaches [36,2,17,44] by adopting a conditional GAN based architecture and training strategy.”
Towards Large-Pose Face Frontalization in the Wild (2017, methodology): “GANs have been used for a wide range of applications including segmentation [33], image generation [4, 5], 3D object generation [27] and so on.”
Generative Adversarial Networks: An Overview (2017, background): “Super-resolution [36], [37], [38] offers an example of how an existing approach can be supplemented with an adversarial loss component to achieve higher quality results.”
Disentangled Representation Learning GAN for Pose-Invariant Face Recognition (2017, methodology): “GAN has been used for image synthesis [6,30], image super resolution [19, 41], and etc.”
NTIRE 2020 Challenge on Real-World Image Super-Resolution: Methods and Results (2020, background): “Recent work has put further emphasis on generating perceptually more appealing predictions using for instance adversarial losses [72, 41, 67].”
Wavelet-SRNet: A Wavelet-Based CNN for Multi-scale Face Super Resolution (2017, methodology): “URDGN [33] promises to predict high-frequency information with an adversarial loss.”
A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications (2020, methodology): “There are methods specific for face SR [317]–[319].”