Filters








6,462 Hits in 3.9 sec

Cascaded Diffusion Models for High Fidelity Image Generation [article]

Jonathan Ho, Chitwan Saharia, William Chan, David J. Fleet, Mohammad Norouzi, Tim Salimans
2021 arXiv   pre-print
We show that cascaded diffusion models are capable of generating high fidelity images on the class-conditional ImageNet generation benchmark, without any assistance from auxiliary image classifiers to  ...  A cascaded diffusion model comprises a pipeline of multiple diffusion models that generate images of increasing resolution, beginning with a standard diffusion model at the lowest resolution, followed  ...  Acknowledgments We thank Jascha Sohl-Dickstein, Douglas Eck and the Google Brain team for feedback, research discussions and technical assistance.  ... 
arXiv:2106.15282v3 fatcat:f3mrzyjv5jcndjdg6icfcs6zrq

Smart, Sparse Contours to Represent and Edit Images [article]

Tali Dekel, Chuang Gan, Dilip Krishnan, Ce Liu, William T. Freeman
2018 arXiv   pre-print
We show that high-quality reconstructions with high fidelity to the source image can be obtained from sparse input, e.g., comprising less than 6% of image pixels.  ...  Our model, based on generative adversarial networks, synthesizes texture and details in regions where no input information is provided.  ...  In this paper, we propose a new method, based on deep generative models, to resolve the conflict between high fidelity and high sparsity.  ... 
arXiv:1712.08232v2 fatcat:p3voddpn6namxiox6js2ylw57u

Sparse, Smart Contours to Represent and Edit Images

Tali Dekel, Chuang Gan, Dilip Krishnan, Ce Liu, William T. Freeman
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
We show that highquality reconstructions with high fidelity to the source image can be obtained from sparse input, e.g., comprising less than 6% of image pixels.  ...  Our model, based on generative adversarial networks, synthesizes texture and details in regions where no input information is provided.  ...  In this paper, we propose a new method, based on deep generative models, to resolve the conflict between high fidelity and high sparsity.  ... 
doi:10.1109/cvpr.2018.00370 dblp:conf/cvpr/DekelGKLF18 fatcat:y623rivpzbbavjb6o47tdqyw6m

Image Super-Resolution via Iterative Refinement [article]

Chitwan Saharia, Jonathan Ho, William Chan, Tim Salimans, David J. Fleet, Mohammad Norouzi
2021 arXiv   pre-print
We further show the effectiveness of SR3 in cascaded image generation, where generative models are chained with super-resolution models, yielding a competitive FID score of 11.3 on ImageNet.  ...  SR3 adapts denoising diffusion probabilistic models to conditional image generation and performs super-resolution through a stochastic denoising process.  ...  We also thank authors of [28] for generously providing us with baseline superresolution samples for human evaluation.  ... 
arXiv:2104.07636v2 fatcat:ae3bac4cyjgg3ayr2gdku2tq3e

The situated laptop: A tangible interface for computer-based studies of surface appearance

B. Darling, J. Ferwerda
2010 Journal of Vision  
, reflectance and illumination properties can be rendered and displayed with great fidelity.  ...  We use this information to drive a physically-based illuminationmap rendering algorithm that generates an accurately oriented and realistically shaded view of a surface to the laptop's display.  ...  System Design • High resolution surface texture Sample experiment to study the influence of diffuse color on the perception of surface gloss.  ... 
doi:10.1167/9.8.324 fatcat:x7yjrhdiencyzmdtzt4odfepge

A Lightweight Optical Flow CNN - Revisiting Data Fidelity and Regularization [article]

Tak-Wai Hui, Xiaoou Tang, Chen Change Loy
2020 arXiv   pre-print
Our network also owns an effective structure for pyramidal feature extraction and embraces feature warping rather than image warping as practiced in FlowNet2 and SPyNet.  ...  It provides high flow estimation accuracy through early correction with seamless incorporation of descriptor matching.  ...  Data Fidelity. Point correspondence across two images is generally constrained by the classical brightness constancy [4] .  ... 
arXiv:1903.07414v3 fatcat:xoftivn44rdsdch7hzzrtg52ou

Signalling ballet in space and time

Boris N. Kholodenko, John F. Hancock, Walter Kolch
2010 Nature reviews. Molecular cell biology  
We thank Mikhail Tsyganov, Javier Muñoz García, Anatoly Kiyatkin and Nikolai Kaimachnikov for discussions.  ...  The short lifetime of the Ras and CD59 clusters is critical for high fidelity signal transmission because it allows for a high sampling rate of the analogue input signal.  ...  Given the similarity between the Ras and GPI-anchored nanocluster systems, it is tempting to speculate that this type of ADA circuitry may represent a general mechanism for high fidelity signal transmission  ... 
doi:10.1038/nrm2901 pmid:20495582 pmcid:PMC2977972 fatcat:x3y744k3hff3la3nyxd6zvdfle

DDRNet: Depth Map Denoising and Refinement for Consumer Depth Cameras Using Cascaded CNNs [chapter]

Shi Yan, Chenglei Wu, Lizhen Wang, Feng Xu, Liang An, Kaiwen Guo, Yebin Liu
2018 Lecture Notes in Computer Science  
We propose a cascaded Depth Denoising and Refinement Network (DDRNet) to tackle this problem by leveraging the multi-frame fused geometry and the accompanying high quality color image through a joint training  ...  Thanks to the well decoupling of the low and high frequency information in the cascaded network, we achieve superior performance over the state-of-the-art techniques.  ...  Although trained on human body dataset, our model also produce high-quality depth map on general objects in arbitrary scenes, eg. the backpack sequence. Fig. 6 . 6 Left: normal map of Din.  ... 
doi:10.1007/978-3-030-01249-6_10 fatcat:ru4ijo7novhj7gfkep5r326s4a

LiteFlowNet: A Lightweight Convolutional Neural Network for Optical Flow Estimation

Tak-Wai Hui, Xiaoou Tang, Chen Change Loy
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
pyramidal feature extraction and embraces feature warping rather than image warping as practiced in FlowNet2.  ...  Our code and trained models are available at github.com/twhui/LiteFlowNet.  ...  This work is supported by SenseTime Group Limited and the General Research Fund sponsored by the Research Grants Council of the Hong Kong SAR (CUHK 14241716, 14224316, 14209217) .  ... 
doi:10.1109/cvpr.2018.00936 dblp:conf/cvpr/HuiTL18 fatcat:pe5o5quuwbgkfedj76xl2dreoa

Global Context with Discrete Diffusion in Vector Quantised Modelling for Image Generation [article]

Minghui Hu, Yujie Wang, Tat-Jen Cham, Jianfei Yang, P.N.Suganthan
2021 arXiv   pre-print
We show that with the help of a content-rich discrete visual codebook from VQ-VAE, the discrete diffusion model can also generate high fidelity images with global context, which compensates for the deficiency  ...  Denoising Diffusion Probabilistic Models (DDPM) in the continuous domain have shown a capability to capture the global context, while generating high-quality images.  ...  In CDM [11] , the authors performed the cascade pipeline on the diffusion model to generate the image with ultra-high and reach state-of-the-art on conditional ImageNet generation.  ... 
arXiv:2112.01799v1 fatcat:tuu4vnww2rcdpaxqvvexbremx4

Stochastic Deep Compressive Sensing for the Reconstruction of Diffusion Tensor Cardiac MRI [chapter]

Jo Schlemper, Guang Yang, Pedro Ferreira, Andrew Scott, Laura-Ann McGill, Zohya Khalique, Margarita Gorodezky, Malte Roehl, Jennifer Keegan, Dudley Pennell, David Firmin, Daniel Rueckert
2018 Lecture Notes in Computer Science  
Our simulation based studies have achieved high reconstruction fidelity and good agreement between DT-CMR parameters obtained with the proposed reconstruction and fully sampled ground truth.  ...  Diffusion Tensor Cardiac Magnetic Resonance (DT-CMR) is a unique non-invasive technique that can resolve the microscopic structure, organisation, and integrity of the myocardium without the need for exogenous  ...  Some of these challenges have been partially addressed or have benefited from parallel imaging techniques for the in-plane acceleration, e.g., using SENSE (sensitivity encoding) and GRAPPA (generalized  ... 
doi:10.1007/978-3-030-00928-1_34 fatcat:bhz5wlc4srgwpocokupi2dvtyu

LiteFlowNet: A Lightweight Convolutional Neural Network for Optical Flow Estimation [article]

Tak-Wai Hui, Xiaoou Tang, Chen Change Loy
2018 arXiv   pre-print
pyramidal feature extraction and embraces feature warping rather than image warping as practiced in FlowNet2.  ...  Our code and trained models are available at https://github.com/twhui/LiteFlowNet .  ...  Given an image pair (I1 and I2), NetC generates two pyramids of high-level features ( Figure 3 : 3 A cascaded flow inference module M :S in NetE.  ... 
arXiv:1805.07036v1 fatcat:2ph5wbhhirco7m52zro5xfl3qe

Discriminative Transfer Learning for General Image Restoration

Lei Xiao, Felix Heide, Wolfgang Heidrich, Bernhard Scholkopf, Michael Hirsch
2018 IEEE Transactions on Image Processing  
In this paper, we propose a discriminative transfer learning method that incorporates formal proximal optimization and discriminative learning for general image restoration.  ...  Recently, several discriminative learning approaches have been proposed for effective image restoration, achieving convincing trade-off between image quality and computational efficiency.  ...  result with different fidelity weight λ TABLE IV : IV Test with different HQS iterations (T ) and model stages (K) for image denoising.  ... 
doi:10.1109/tip.2018.2831925 pmid:29993740 fatcat:prsa74c75jhyhnufmzbkykhqza

FastDiff: A Fast Conditional Diffusion Model for High-Quality Speech Synthesis [article]

Rongjie Huang, Max W. Y. Lam, Jun Wang, Dan Su, Dong Yu, Yi Ren, Zhou Zhao
2022 arXiv   pre-print
This paper proposes FastDiff, a fast conditional diffusion model for high-quality speech synthesis.  ...  Denoising diffusion probabilistic models (DDPMs) have recently achieved leading performances in many generative tasks.  ...  generalization, indicating that FastDiff could universally generate high-fidelity audio from entirely new (unseen) speakers outside the train set.  ... 
arXiv:2204.09934v1 fatcat:mozdu3jn3ngv3ms2jr7zuuftaa

Diffusion-Steered Super-Resolution Image Reconstruction [chapter]

Baraka J. Maiseli
2018 Colorimetry and Image Processing  
For decades, super-resolution has been a widely applied technique to improve the spatial resolution of an image without hardware modification.  ...  Extensive analysis of the proposed resolution-enhancement model shows that it can respond well on different image regions.  ...  , which exploits information from a sequence of degraded images to generate a high-quality image [2, 6] .  ... 
doi:10.5772/intechopen.71024 fatcat:yogrp3q4j5h35osvb3z3r7px4i
« Previous Showing results 1 — 15 out of 6,462 results