166 Hits in 5.6 sec

Face Super-Resolution Guided by Facial Component Heatmaps [chapter]

Xin Yu, Basura Fernando, Bernard Ghanem, Fatih Porikli, Richard Hartley
2018 Lecture Notes in Computer Science  
Our CNN has two branches: one for super-resolving face images and the other branch for predicting salient regions of a face coined facial component heatmaps.  ...  State-of-the-art face super-resolution methods leverage deep convolutional neural networks to learn a mapping between low-resolution (LR) facial patterns and their corresponding high-resolution (HR) counterparts  ...  With the help of our facial component heatmap estimation branch, our method super-resolves faces in different poses without distortions caused by erroneous facial component localization in LR inputs.  ... 
doi:10.1007/978-3-030-01240-3_14 fatcat:v6tjecsnz5ctraypkckjt3jzfe

Face Super-Resolution Guided by 3D Facial Priors [article]

Xiaobin Hu, Wenqi Ren, John LaMaster, Xiaochun Cao, Xiaoming Li, Zechao Li, Bjoern Menze, Wei Liu
2020 arXiv   pre-print
State-of-the-art face super-resolution methods employ deep convolutional neural networks to learn a mapping between low- and high- resolution facial patterns by exploring local appearance knowledge.  ...  In this paper, we propose a novel face super-resolution method that explicitly incorporates 3D facial priors which grasp the sharp facial structures.  ...  The bottom block is dedicated to face super-resolution guided by the facial coefficients and rendered sharp face structures which are concatenated by the Spatial Feature Transform (SFT) layer.  ... 
arXiv:2007.09454v1 fatcat:gyxdruuling7bdaoetiwjkigei

Deep Learning-based Face Super-Resolution: A Survey [article]

Junjun Jiang, Chenyang Wang, Xianming Liu, Jiayi Ma
2021 arXiv   pre-print
Face super-resolution (FSR), also known as face hallucination, which is aimed at enhancing the resolution of low-resolution (LR) face images to generate high-resolution (HR) face images, is a domain-specific  ...  A curated list of papers and resources to face super-resolution are available at  ...  Later on, super-resolution guided by 3D facial priors (FSRG3DFP) [120] estimates 3D priors instead of 2D priors to learn 3D facial details and capture facial component information by the spatial feature  ... 
arXiv:2101.03749v2 fatcat:q56d2mpn4rfyzmi5fo36d2ecja

Deep Learning-based Face Super-resolution: A Survey

Junjun Jiang, Chenyang Wang, Xianming Liu, Jiayi Ma
2023 ACM Computing Surveys  
Face super-resolution (FSR), also known as face hallucination, which is aimed at enhancing the resolution of low-resolution (LR) face images to generate high-resolution face images, is a domain-specific  ...  image super-resolution problem.  ...  Later, super-resolution guided by three-dimensional (3D) facial priors (FSRG3DFP) [120] estimates 3D priors instead of 2D priors to learn 3D facial details and capture facial component information by  ... 
doi:10.1145/3485132 fatcat:gxaod3eqs5gkbmtow3k4jrrpvy

Component Attention Guided Face Super-Resolution Network: CAGFace [article]

Ratheesh Kalarot, Tao Li, Fatih Porikli
2019 arXiv   pre-print
multi-stage neural network for 4× super-resolution for face images.  ...  We implicitly impose facial component-wise attention maps using a segmentation network to allow our network to focus on face-inherent patterns.  ...  They do not use facial components to guide the super-resolution process. Most use lower resolution images in their training, which may be further limiting their representation capacity.  ... 
arXiv:1910.08761v1 fatcat:z6nymdxp3fb25gmoiiljvqp5c4

Deep Face Super-Resolution with Iterative Collaboration between Attentive Recovery and Landmark Estimation [article]

Cheng Ma, Zhenyu Jiang, Yongming Rao, Jiwen Lu, Jie Zhou
2020 arXiv   pre-print
However, the prior knowledge is not fully exploited in existing methods, since facial priors such as landmark and component maps are always estimated by low-resolution or coarsely super-resolved images  ...  In this paper, we propose a deep face super-resolution (FSR) method with iterative collaboration between two recurrent networks which focus on facial image recovery and landmark estimation respectively  ...  Facial component heatmaps are predicted in [39] to provide localizations of facial components for improving the SR quality.  ... 
arXiv:2003.13063v1 fatcat:ivo5477i4fffzbyk44j2yqjrwi

Edge and Identity Preserving Network for Face Super-Resolution [article]

Jonghyun Kim, Gen Li, Inyong Yun, Cheolkon Jung, Joongkyu Kim
2020 arXiv   pre-print
Face super-resolution has become an indispensable part in security problems such as video surveillance and identification system, but the distortion in facial components is a main obstacle to overcoming  ...  The proposed methods facilitate our super-resolution network to elaborately restore facial components and generate enhanced 8x scaled super-resolution images with a lightweight network structure.  ...  which overcomes distortion of facial components by providing the edge information and data distributions in the face super-resolution process.  ... 
arXiv:2008.11977v1 fatcat:tvjcpnbmbzd2nbpakhksvkotwy

FaceFormer: Scale-aware Blind Face Restoration with Transformers [article]

Aijin Li, Gen Li, Lei Sun, Xintao Wang
2022 arXiv   pre-print
Thus, our FaceFormer achieves fidelity and robustness restored faces, which possess realistic and symmetrical details of facial components.  ...  In this work, we propose a novel scale-aware blind face restoration framework, named FaceFormer, which formulates facial feature restoration as scale-aware transformation.  ...  FSRNet [11] employs face landmark heatmaps and parsing maps to refine the super-resolution results, especially for misaligned LR faces.  ... 
arXiv:2207.09790v1 fatcat:23wajwzjszashnaqkvrbz32dky

GFNet: A Gradient Information Compensation-based Face Super-Resolution Network

Shengxaing Luo, Jinbo Lu
2022 IEEE Access  
Face super-resolution (FSR) is defined as the generation of high-resolution face images from low-resolution face images.  ...  However, the additional data requires manual labeling, and facial landmark heatmaps and parsing maps cannot represent the intrinsic geometric structure of facial components.  ...  ; 2) estimating face priors from low-resolution inputs is itself a difficult task; and 3) facial landmark heatmaps and parsing maps cannot represent the intrinsic geometric structure of facial components  ... 
doi:10.1109/access.2022.3143499 fatcat:xn5y6cpzfjhdnixn7257k2ezxi

Multiple Exemplars-based Hallucinationfor Face Super-resolution and Editing [article]

Kaili Wang, Jose Oramas, Tinne Tuytelaars
2021 arXiv   pre-print
Besides standard face super-resolution, our method allows to perform subtle face editing simply by replacing the exemplars with another set with different facial features.  ...  Given a really low-resolution input image of a face (say 16x16 or 8x8 pixels), the goal of this paper is to reconstruct a high-resolution version thereof.  ...  Conclusion In this paper, we propose to use multiple exemplars as conditions to guide the model to super-resolve LR images.  ... 
arXiv:2009.07827v3 fatcat:zvs5xharw5aw7oybjmxgibzqam

Multi-Prior Learning via Neural Architecture Search for Blind Face Restoration [article]

Yanjiang Yu, Puyang Zhang, Kaihao Zhang, Wenhan Luo, Changsheng Li, Ye Yuan, Guoren Wang
2022 arXiv   pre-print
Blind Face Restoration (BFR) aims to recover high-quality face images from low-quality ones and usually resorts to facial priors for improving restoration performance.  ...  In this way, MFPSNet takes full advantage of semantic-level (parsing maps), geometric-level (facial heatmaps), reference-level (facial dictionaries) and pixel-level (degraded images) information and thus  ...  [39] propose their methods based on facial heatmaps. Chen et al. [5] develop a coarse-to-fine network and apply facial component heatmaps to their fine super-resolution network.  ... 
arXiv:2206.13962v1 fatcat:mkwpisnkwfahtcfhvq4jciypsq

Pro-UIGAN: Progressive Face Hallucination from Occluded Thumbnails [article]

Yang Zhang, Xin Yu, Xiaobo Lu, Ping Liu
2022 arXiv   pre-print
Pro-UIGAN iteratively (1) estimates facial geometry priors for low-resolution (LR) faces and (2) acquires non-occluded HR face images under the guidance of the estimated priors.  ...  In this paper, we study the task of hallucinating an authentic high-resolution (HR) face from an occluded thumbnail.  ...  During this procedure, the facial geometry priors provide spatial configuration of facial components and shapes, guiding joint face inpainting and super-resolution.  ... 
arXiv:2108.00602v6 fatcat:o7pkr3vbxnbqlis5cqucoy2hlq

Frequency Aware Face Hallucination Generative Adversarial Network with Semantic Structural Constraint [article]

Shailza Sharma, Abhinav Dhall, Vinay Kumar
2021 arXiv   pre-print
Most current face hallucination methods rely on two-dimensional facial priors to generate high resolution face images from low resolution face images.  ...  To explicitly encode the high frequency components, an auto encoder is proposed to generate high resolution coefficients of Discrete Cosine Transform (DCT).  ...  Two CNN branches-one with the facial structural information (alligned heatmaps of nose, eyes, skin and chin) and the other for the face super resolution are aggregated by Yu et al. [9] .  ... 
arXiv:2110.01880v1 fatcat:chpwmrm34vcq5p3moeuo3gzudu

Component Semantic Prior Guided Generative Adversarial Network for Face Super-Resolution

Lu Liu, Shenghui Wang, Lili Wan
2019 IEEE Access  
INDEX TERMS Facial component, face super-resolution, generative adversarial networks, multiple task, semantic prior.  ...  Face super-resolved (SR) images aid human perception. The state-of-the-art face SR methods leverage the spatial location of facial components as prior knowledge.  ...  INTRODUCTION Face super-resolution (SR) task refers to reconstructing a high-resolution (HR) image from a low-resolution (LR) facial image.  ... 
doi:10.1109/access.2019.2921859 fatcat:sz3t7jjdwzh2fmtfnvs2lniu7e

Facial image super-resolution guided by adaptive geometric features

Zhenfeng Fan, Xiyuan Hu, Chen Chen, Xiaolian Wang, Silong Peng
2020 EURASIP Journal on Wireless Communications and Networking  
This paper addresses the traditional issue of restoring a high-resolution (HR) facial image from a low-resolution (LR) counterpart.  ...  As a special case of general images, the face has limited geometric variations, which we believe that the relevant depth map can be learned and used to guide the face SR task.  ...  Acknowledgements This work is supported in part by the National Key R&D Program of China (2017YFC0803505) and the Natural Science Foundation of China (61571438).  ... 
doi:10.1186/s13638-020-01760-y fatcat:7rhc4uvtwvadtcextzapvqb6re
« Previous Showing results 1 — 15 out of 166 results