438 Hits in 6.0 sec

Resolution-robust Large Mask Inpainting with Fourier Convolutions [article]

Roman Suvorov, Elizaveta Logacheva, Anton Mashikhin, Anastasia Remizova, Arsenii Ashukha, Aleksei Silvestrov, Naejin Kong, Harshith Goka, Kiwoong Park, Victor Lempitsky
2021 arXiv   pre-print
Modern image inpainting systems, despite the significant progress, often struggle with large missing areas, complex geometric structures, and high-resolution images.  ...  LaMa is based on i) a new inpainting network architecture that uses fast Fourier convolutions (FFCs), which have the image-wide receptive field; ii) a high receptive field perceptual loss; iii) large training  ...  LaMa can capture and generate complex periodic structures, and is robust to large masks.  ... 
arXiv:2109.07161v2 fatcat:eawdlov2inhyldqonov7sudfwq

Keys to Better Image Inpainting: Structure and Texture Go Hand in Hand [article]

Jitesh Jain, Yuqian Zhou, Ning Yu, Humphrey Shi
2022 arXiv   pre-print
Deep image inpainting has made impressive progress with recent advances in image generation and processing algorithms.  ...  convolutional layers.  ...  The image generation ability associated with StyleGAN2 from stochasticity enables the large-hole filling with realistic structures even when the global context lacks. Large Mask Inpainting.  ... 
arXiv:2208.03382v1 fatcat:yqpxx2fqazflxeymqa7awlupgi

GLaMa: Joint Spatial and Frequency Loss for General Image Inpainting [article]

Zeyu Lu and Junjun Jiang and Junqin Huang and Gang Wu and Xianming Liu
2022 arXiv   pre-print
By incorporating more degraded images in the training phase, we can expect to enhance the robustness of the model with respect to various masks.  ...  In recent years, thanks to the resurgence of convolutional neural networks (CNNs), image inpainting task has made great breakthroughs.  ...  It can be seen from Figs. 1 and 7 that LaMa [31] is not good at repairing the large-area mask of high-resolution face.  ... 
arXiv:2205.07162v1 fatcat:x22uprxnsngozn5twsxx2oph5q

Incremental Transformer Structure Enhanced Image Inpainting with Masking Positional Encoding [article]

Qiaole Dong, Chenjie Cao, Yanwei Fu
2022 arXiv   pre-print
Furthermore, a masking positional encoding strategy is utilized to improve the performance with large irregular masks.  ...  The proposed model restores holistic image structures with a powerful attention-based transformer model in a fixed low-resolution sketch space.  ...  [45] utilize the Fast Fourier Convolution (FFC) to encode features in frequency fields with global receptive fields for resolution-robust inpainting.  ... 
arXiv:2203.00867v2 fatcat:xehfr7t4grdulky5ebqst4hmxu

CM-GAN: Image Inpainting with Cascaded Modulation GAN and Object-Aware Training [article]

Haitian Zheng, Zhe Lin, Jingwan Lu, Scott Cohen, Eli Shechtman, Connelly Barnes, Jianming Zhang, Ning Xu, Sohrab Amirghodsi, Jiebo Luo
2022 arXiv   pre-print
Recent image inpainting methods have made great progress but often struggle to generate plausible image structures when dealing with large holes in complex images.  ...  We propose cascaded modulation GAN (CM-GAN), a new network design consisting of an encoder with Fourier convolution blocks that extract multi-scale feature representations from the input image with holes  ...  Our contributions are four-fold: -Cascaded Modulation GAN, a new inpainting network architecture formed by a masked image encoder with Fourier convolution blocks and a cascaded globalspatial modulation-based  ... 
arXiv:2203.11947v3 fatcat:vnfyuua5oracvmc27h4hahlo3q

Towards Unified Keyframe Propagation Models [article]

Patrick Esser and Peter Michael and Soumyadip Sengupta
2022 arXiv   pre-print
The global interaction stream remains robust in difficult situations such as large camera motions, where explicit alignment fails.  ...  Applied to video inpainting, our approach leads to 44% and 26% improvements in FID and LPIPS scores. Code at  ...  We use the 'flickr-all' video subset combined with the 'fvi-fgs-h' mask subset which contains large inpainting masks and represents a challenging setting.  ... 
arXiv:2205.09731v1 fatcat:sbkcvbeppzay5goizpoextxw74

Image Inpainting: A Review

Omar Elharrouss, Noor Almaadeed, Somaya Al-Maadeed, Younes Akbari
2019 Neural Processing Letters  
With the improvement of image processing tools and the flexibility of digital image editing, automatic image inpainting has found important applications in computer vision and has also become an important  ...  The main contribution of this paper is the presentation of the three categories of image inpainting methods along with a list of available datasets that the researchers can use to evaluate their proposed  ...  The images resolution is defined by three categories full-size with 1240 1110 pixels, half size with 690 555 pixels and the third resolution with 413 370.  ... 
doi:10.1007/s11063-019-10163-0 fatcat:b4n4h4ts7bchfl5s5kjikq4npy

Architecture-Agnostic Masked Image Modeling – From ViT back to CNN [article]

Siyuan Li, Di Wu, Fang Wu, Zelin Zang, Kai Wang, Lei Shang, Baigui Sun, Hao Li, Stan.Z.Li
2022 arXiv   pre-print
Masked image modeling (MIM), an emerging self-supervised pre-training method, has shown impressive success across numerous downstream vision tasks with Vision transformers (ViT).  ...  Based on this fact, we propose an Architecture-Agnostic Masked Image Modeling framework (A^2MIM), which is compatible with not only Transformers but also CNNs in a unified way.  ...  For CV, inpainting tasks [36] to predict large missing regions using convolutional networks is proposed to learn representation without supervision.  ... 
arXiv:2205.13943v2 fatcat:dq6pmnfqhbbrjiv42auuoqdday

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
Then we employ a large-mask inpainting method with fast Fourier convolutions to remove the detected distracting objects.  ...  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  ...  In [32] , robust Large Mask (LaMa) inpainting with Fourier convolutions was developed with resolution, which will serve as the baseline for our research.  ... 
arXiv:2205.06934v1 fatcat:6lv7hv7zk5e6vczlxtord4rqbi

FD-MAR: Fourier Dual-domain Network for CT Metal Artifact Reduction [article]

Zilong Li, Qi Gao, Yaping Wu, Chuang Niu, Junping Zhang, Meiyun Wang, Ge Wang, Hongming Shan
2022 arXiv   pre-print
To address these issues, this paper explores fast Fourier convolution for MAR in both sinogram and image domains, and proposes a Fourier dual-domain network for MAR, termed FD-MAR.  ...  Specifically, we first propose a Fourier sinogram restoration network, which can leverage sinogram-wide receptive context to fill in the metal-corrupted region from uncorrupted region and, hence, is robust  ...  We use Fast Fourier convolution [9] to perform the global inpainting in FS-Net.  ... 
arXiv:2207.11678v1 fatcat:2xantmw4xncq3oz7qutcwnyqeq

Comparison of CoModGANs, LaMa and GLIDE for Art Inpainting- Completing M.C Escher's Print Gallery [article]

Lucia Cipolina-Kun and Simone Caenazzo and Gaston Mazzei
2022 arXiv   pre-print
This work compares three current state-of-the art models for inpainting of large missing regions.  ...  Digital art restoration has benefited from inpainting models to correct the degradation or missing sections of a painting.  ...  The model Resolution-robust Large Mask Inpainting with Fourier Convolutions was designed with large regions in mind as well.  ... 
arXiv:2205.01741v1 fatcat:3ianeyhgznf6jnolcdcxx6gshe

Overview of Image Inpainting and Forensic Technology

Kai Liu, Junke Li, Syed Sabahat Hussain Bukhari
2022 Security and Communication Networks  
With the progress of image processing theory, digital image inpainting technology has also developed rapidly. It not only enhances the image expression but also provides tools for image forgery.  ...  This article reviews the development of image inpainting and forensic technology and systematically summarizes and scientifically classifies the current research work.  ...  Stable performance; suitable for image inpainting with irregular shape. 19.04 dB (PSNR) Deep CNN Gated convolution High-resolution Using gated convolution to optimize partial convolution Content network  ... 
doi:10.1155/2022/9291971 doaj:9085f6d15bc64d13acd70f636181bf23 fatcat:pg5ncxzngjblzl42bstp3ykxpi

A Novel Approach for Video Inpainting Using Autoencoders

Irfan Siddavatam, Department of Information Technology, K J Somaiya College of Engineering, Mumbai, Maharashtra, India, Ashwini Dalvi, Dipti Pawade, Akshay Bhatt, Jyeshtha Vartak, Arnav Gupta
2021 International Journal of Information Engineering and Electronic Business  
Inpainting is a task undertaken to fill in damaged or missing parts of an image or video frame, with believable content.  ...  Video Inpainting, an extension of Image Inpainting, is a much more challenging task due to the constraint added by the time dimension.  ...  Frame-Recurrent Video Inpainting by Robust Optical Flow Inference [22] Free-form Video Inpainting with 3D Gated Convolution and Objectives  The generator uses a 3D convolutional layer to service the masked  ... 
doi:10.5815/ijieeb.2021.06.05 fatcat:d3ajmlh4l5dthjofbpetghsely

RePaint: Inpainting using Denoising Diffusion Probabilistic Models [article]

Andreas Lugmayr, Martin Danelljan, Andres Romero, Fisher Yu, Radu Timofte, Luc Van Gool
2022 arXiv   pre-print
Free-form inpainting is the task of adding new content to an image in the regions specified by an arbitrary binary mask.  ...  We validate our method for both faces and general-purpose image inpainting using standard and extreme masks.  ...  ] to increase the receptive field, Partial Convolutions [19] and Gated Convolutions [48] to guide the convolution kernel according to the inpainted mask, Contextual Attention [46] to leverage on  ... 
arXiv:2201.09865v4 fatcat:5p2suiozkbcixpbpk2ohggt4jy

A hybrid approach to cosmic microwave background lensing reconstruction from all-sky intensity maps

S. Plaszczynski, A. Lavabre, L. Perotto, J.-L. Starck
2012 Astronomy and Astrophysics  
It implies using a large galactic mask and dealing with a strong inhomogeneous noise.  ...  For l < 100, we show that a full-sky inpainting method, already described in a previous work, still allows a minimal variance reconstruction, with a bias that must be accounted for by a Monte-Carlo method  ...  Among the inpainting algorithm implemented within the multi-resolution on the sphere (MRS) package 4 , we found that the most robust results are obtained with the spherical harmonics L 1 norm minimization  ... 
doi:10.1051/0004-6361/201218899 fatcat:5pwydv7t6faltkw66iopcy26hu
« Previous Showing results 1 — 15 out of 438 results