298 Hits in 2.4 sec

Fully Context-Aware Image Inpainting with a Learned Semantic Pyramid [article]

Wendong Zhang, Yunbo Wang, Junwei Zhu, Ying Tai, Bingbing Ni, Xiaokang Yang
2021 arXiv   pre-print
Within the prior learner, we present an optional module for variational inference to realize probabilistic image inpainting driven by various learned priors.  ...  Although recent image inpainting models have made significant progress in generating vivid visual details, they can still lead to texture blurring or structural distortions due to contextual ambiguity  ...  in image inpainting.  ... 
arXiv:2112.04107v1 fatcat:2zyu5jabcncxdlnicxylcaayym

Inpainting Digital Dunhuang Murals with Structure-Guided Deep Network

Zhiheng Zhou, Xinran Liu, Junyuan Shang, Junchu Huang, Zhihao Li, Haiping Jia
2022 ACM Journal on Computing and Cultural Heritage  
In this paper, we propose a deep-learning-based structure-guided inpainting method for Dunhuang mural image, which utilizes relevant color information in deep features to improve the color inpainting quality  ...  Algorithms of mural image inpainting help simplify the digital restoration process of the deteriorated murals.  ...  Distillation-guided [37] applies direct feature level supervision during network training, guiding the encoder to build embedding of the missing region, which is a way to implicitly reine features from  ... 
doi:10.1145/3532867 fatcat:ghi44gvaqjhdfa7t2ulhpkaeji

Generative Memory-Guided Semantic Reasoning Model for Image Inpainting [article]

Xin Feng, Wenjie Pei, Fengjun Li, Fanglin Chen, David Zhang, Guangming Lu
2022 arXiv   pre-print
In this paper, we propose the Generative Memory-Guided Semantic Reasoning Model (GM-SRM), which not only learns the intra-image priors from the known regions, but also distills the inter-image reasoning  ...  Most existing methods for image inpainting focus on learning the intra-image priors from the known regions of the current input image to infer the content of the corrupted regions in the same image.  ...  Generative prior-guided inpainting methods. Learning the distribution of semantic patterns by generative models leads to promising performance on image inpainting.  ... 
arXiv:2110.00261v2 fatcat:jpkvxxe4wberja5tankbba23c4

A Saliency-Guided Street View Image Inpainting Framework for Efficient Last-Meters Wayfinding [article]

Chuanbo Hu, Shan Jia, Fan Zhang, Xin Li
2022 arXiv   pre-print
To address this problem, we highlight the importance of reducing visual distraction in image-based wayfinding by proposing a saliency-guided image inpainting framework.  ...  Experimental results with both qualitative and quantitative analysis show that our saliency-guided inpainting method can not only achieve great perceptual quality in street view images but also redirect  ...  Section 3 describes the proposed saliency-guided street view image inpainting approach.  ... 
arXiv:2205.06934v1 fatcat:6lv7hv7zk5e6vczlxtord4rqbi

Occlusion-Aware Video Object Inpainting [article]

Lei Ke, Yu-Wing Tai, Chi-Keung Tang
2021 arXiv   pre-print
Conventional video inpainting is neither object-oriented nor occlusion-aware, making it liable to obvious artifacts when large occluded object regions are inpainted.  ...  This paper presents occlusion-aware video object inpainting, which recovers both the complete shape and appearance for occluded objects in videos given their visible mask segmentation.  ...  Modal Masks After Flow-guided pixel propagation Video Object Inpainting Results Image Frame + Complete Object Masks (c) Flow-guided Video Object Inpainting concat !  ... 
arXiv:2108.06765v1 fatcat:dltn4glbjvfq7haccbhhmsznpm

ReGO: Reference-Guided Outpainting for Scenery Image [article]

Yaxiong Wang, Yunchao Wei, Xueming Qian, Li Zhu, Yi Yang
2022 arXiv   pre-print
To overcome the weakness, this work investigates a principle way to synthesize texture-rich results by borrowing pixels from its neighbors (i.e., reference images), named Reference-Guided Outpainting (  ...  Recently, generative adversarial learning has significantly advanced the image outpainting by producing semantic consistent content for the given image.  ...  Image-Guided Convolution The proposed Image-Guided Convolution (IGConv) aims to help the network compensate texture details for the new content using the distilled the beneficial features from the reference  ... 
arXiv:2106.10601v4 fatcat:n3hhc6gl2rdxxkg4ftw4jncvj4

View Blind-spot as Inpainting: Self-Supervised Denoising with Mask Guided Residual Convolution [article]

Yuhongze Zhou, Liguang Zhou, Tin Lun Lam, Yangsheng Xu
2021 arXiv   pre-print
Therefore, we propose a novel Mask Guided Residual Convolution (MGRConv) into common convolutional neural networks, e.g. U-Net, to promote blind-spot based denoising.  ...  In this paper, we take an intuitive view of blind-spot strategy and consider its process of using neighbor pixels to predict manipulated pixels as an inpainting process.  ...  by itself automatically, which produces higher-quality traditional and user-guided inpainting results.  ... 
arXiv:2109.04970v1 fatcat:oh5vrygglfasphtp46rbq6s5my

Learning Prior Feature and Attention Enhanced Image Inpainting [article]

Chenjie Cao, Qiaole Dong, Yanwei Fu
2022 arXiv   pre-print
Many recent inpainting works have achieved impressive results by leveraging Deep Neural Networks (DNNs) to model various prior information for image restoration.  ...  However, it is nontrivial to directly replace the new backbones in inpainting networks, as the inpainting is an inverse problem fundamentally different from the recognition tasks.  ...  Then, features from the MAE decoder are added to the inpainting CNN for the prior guided image inpainting.  ... 
arXiv:2208.01837v1 fatcat:jpqv2p4zanf3deduxa5u6kawea

CR-Fill: Generative Image Inpainting with Auxiliary Contexutal Reconstruction [article]

Yu Zeng, Zhe Lin, Huchuan Lu, Vishal M. Patel
2021 arXiv   pre-print
CR loss) is required only during training, and only the inpainting generator is required during the inference.  ...  Recent deep generative inpainting methods use attention layers to allow the generator to explicitly borrow feature patches from the known region to complete a missing region.  ...  CR loss can be extended for distilling more information to an inpainting model through the reconstruction of higher-level represen-tation of images, e.g. semantic layout.  ... 
arXiv:2011.12836v2 fatcat:sfnhw7c6ajfk3h5cc4c5xqk6fi

Information-Theoretic Segmentation by Inpainting Error Maximization [article]

Pedro Savarese and Sunnie S. Y. Kim and Michael Maire and Greg Shakhnarovich and David McAllester
2021 arXiv   pre-print
An easily computed loss drives a greedy search process to maximize inpainting error over these partitions.  ...  We study image segmentation from an information-theoretic perspective, proposing a novel adversarial method that performs unsupervised segmentation by partitioning images into maximally independent sets  ...  Left to right: input image; ground truth mask; IEM mask; inpainting result, with every pixel inpainted as FG or BG according to the IEM mask; SegNet mask Image Ground-truth IEM Result Inpainted Image  ... 
arXiv:2012.07287v3 fatcat:lvojymt5zzeoddjjvfijrxalfu

Few-Shot Head Swapping in the Wild [article]

Changyong Shu, Hemao Wu, Hang Zhou, Jiaming Liu, Zhibin Hong, Changxing Ding, Junyu Han, Jingtuo Liu, Errui Ding, Jingdong Wang
2022 arXiv   pre-print
Particularly, seamless blending is achieved with the help of a Semantic-Guided Color Reference Creation procedure and a Blending UNet.  ...  I BG T = I T * (1.0 − M H A ) is the background image without the target head. Semantic-Guided Color Reference Creation.  ...  Semantic-Guided Color Reference Creation.  ... 
arXiv:2204.13100v1 fatcat:nosfg6s6evcjxl4zxwljqkfoim

Progressively Inpainting Images Based on a Forked-Then-Fused Decoder Network

Shuai Yang, Rong Huang, Fang Han
2021 Sensors  
Image inpainting aims to fill in corrupted regions with visually realistic and semantically plausible contents.  ...  In this paper, we propose a progressive image inpainting method, which is based on a forked-then-fused decoder network.  ...  The ill-posedness of image inpainting can be distilled into the following: how to seek the most proper hypothesis for the corrupted region conditioned on the valid surroundings.  ... 
doi:10.3390/s21196336 pmid:34640656 pmcid:PMC8512423 fatcat:oxhextdavvg2regfz3d4redxnm

Multi-phase Liver Tumor Segmentation with Spatial Aggregation and Uncertain Region Inpainting [article]

Yue Zhang, Chengtao Peng, Liying Peng, Huimin Huang, Ruofeng Tong, Lanfen Lin, Jingsong Li, Yen-Wei Chen, Qingqing Chen, Hongjie Hu, Zhiyi Peng
2021 arXiv   pre-print
Multi-phase computed tomography (CT) images provide crucial complementary information for accurate liver tumor segmentation (LiTS).  ...  Moreover, we devise an uncertain region inpainting module (URIM) to refine uncertain pixels using neighboring discriminative features.  ...  During the refinement stage, uncertain pixels gradually absorb more distant confident features while the uncertain regions shrink, which works like the image inpainting [9, 21] .  ... 
arXiv:2108.00911v2 fatcat:kbiklrj7ebdzvm6bjk4dy5avt4

GLocal: Global Graph Reasoning and Local Structure Transfer for Person Image Generation [article]

Liyuan Ma, Kejie Huang, Dongxu Wei, Haibin Shen
2021 arXiv   pre-print
in texture inpainting.  ...  In this paper, we focus on person image generation, namely, generating person image under various conditions, e.g., corrupted texture or different pose.  ...  Thus we introduce following learning objectives to guide the training process. Pixel-wise Loss.  ... 
arXiv:2112.00263v1 fatcat:q23q35rvavdg7lliopdesj2gqm

Exploiting the Inherent Limitation of L0 Adversarial Examples [article]

Fei Zuo, Bokai Yang, Xiaopeng Li, Lannan Luo, Qiang Zeng
2019 arXiv   pre-print
More concretely, given an image I, it is pre-processed to obtain another image I' .  ...  In addition, we show that the pre-processing technique, inpainting, used for detection can also work as an effective defense, which has a high probability of removing the adversarial influence of L0 perturbations  ...  Specifically, we propose an inpainting-based algorithm to process images, where inpainting refers to the process of reconstructing the lost or corrupted parts of an image.  ... 
arXiv:1812.09638v3 fatcat:35tmtvgjyjd3zpep5d3a3q5age
« Previous Showing results 1 — 15 out of 298 results