Filters








276 Hits in 4.5 sec

Patch-based blind deconvolution with parametric interpolation of convolution kernels

Filip Sroubek, Michal Sorel, Irena Horackova, Jan Flusser
2013 2013 IEEE International Conference on Image Processing  
Space-variant deconvolution of photos blurred by real camera shake: (a) one input blurry image; (b) reconstruction using the proposed method with parametric blur interpolation; (c) close-ups: (left) input  ...  Blurs are first estimated in a small number of image patches.  ...  The SV convolution model in (1) transforms into a set of convolutions g p = k p * u p , (2) where k p is a convolution kernel that approximates h(s, t, x, y) for (x, y) in the p-th patch.  ... 
doi:10.1109/icip.2013.6738119 dblp:conf/icip/SroubekSHF13 fatcat:2vun5ubblvhevnpb5i6igqzdvu

Semi-Blind Spatially-Variant Deconvolution in Optical Microscopy with Local Point Spread Function Estimation by Use of Convolutional Neural Networks

Adrian Shajkofci, Michael Liebling
2018 2018 25th IEEE International Conference on Image Processing (ICIP)  
To find the local PSF map in a computationally tractable way, we train a convolutional neural network to perform regression of an optical parametric model on synthetically blurred image patches.  ...  We present a semi-blind, spatially-variant deconvolution technique aimed at optical microscopy that combines a local estimation step of the point spread function (PSF) and deconvolution using a spatially  ...  to the patch m with parameters a m and M the total number of patches and thus local PSF kernels.  ... 
doi:10.1109/icip.2018.8451736 dblp:conf/icip/ShajkofciL18 fatcat:apvlnr7nufbctiapvyz3fz3byu

Self-Calibration of Optical Lenses

Michael Hirsch, Bernhard Scholkopf
2015 2015 IEEE International Conference on Computer Vision (ICCV)  
We propose a method that enables the self-calibration of lenses from a natural image, or a set of such images.  ...  To this end we develop a machine learning framework that is able to exploit several recorded images and distills the available information into an accurate model of the considered lens.  ...  With the PSF model at hand, we correct the images captured with it via non-blind nonstationary deconvolution.  ... 
doi:10.1109/iccv.2015.77 dblp:conf/iccv/HirschS15 fatcat:bvex5tdb5bgadoztz3onksooq4

Space-Variant Single-Image Blind Deconvolution for Removing Camera Shake

Stefan Harmeling, Michael Hirsch, Bernhard Schölkopf
2010 Neural Information Processing Systems  
for single image blind deconvolution for space-invariant filters by Cho and Lee to construct a method for blind deconvolution in the case of space-variant blur, and (iii), we present an experimental setup  ...  Modelling camera shake as a space-invariant convolution simplifies the problem of removing camera shake, but often insufficiently models actual motion blur such as those due to camera rotation and movements  ...  Blind deconvolution with smoothly varying PSF We now outline a single image blind deconvolution algorithm for space-variant blur, generalizing the method of Cho and Lee [2] , that aims to recover a sharp  ... 
dblp:conf/nips/HarmelingHS10 fatcat:3s77ixyerfedtf2v2755hu7vga

Deblurring by Example Using Dense Correspondence

Yoav Hacohen, Eli Shechtman, Dani Lischinski
2013 2013 IEEE International Conference on Computer Vision  
First, we suggest exploiting information from the reference image not only for blur kernel estimation, but also as a powerful local prior for the non-blind deconvolution step.  ...  Unlike the above previous method, which has proven successful only with simple deblurring scenarios, we demonstrate that our method succeeds on a variety of real-world examples.  ...  We found that any small interpolation/upsampling error in the kernel might result in large deconvolution artifacts.  ... 
doi:10.1109/iccv.2013.296 dblp:conf/iccv/HaCohenSL13 fatcat:towbhmbxivavlfivtvsxjjmxfe

Spatially-Variant CNN-based Point Spread Function Estimation for Blind Deconvolution and Depth Estimation in Optical Microscopy

Adrian Shajkofci, Michael Liebling
2020 IEEE Transactions on Image Processing  
When the recovered PSFs are used with a spatially-variant and regularized Richardson-Lucy (RL) deconvolution algorithm, we observed up to 2.1 dB better Signal-to-Noise Ratio (SNR) compared to other Blind  ...  Deconvolution (BD) techniques.  ...  Many blind deconvolution techniques are computationally expensive, especially for larger convolution kernels, and assume spatially invariant PSFs.  ... 
doi:10.1109/tip.2020.2986880 pmid:32305918 fatcat:4gkiszajobccdo6wuj2li22vcu

Crowd Counting by Adapting Convolutional Neural Networks with Side Information [article]

Di Kang, Debarun Dhar, Antoni B. Chan
2016 arXiv   pre-print
In particular, we model the filter weights as a low-dimensional manifold, parametrized by the side information, within the high-dimensional space of filter weights.  ...  On experiments in crowd counting, the ACNN improves counting accuracy compared to a plain CNN with a similar number of parameters.  ...  available side information, and validate it on various crowd counting datasets and a non-blind image deconvolution task; 2) Using ACNN, we achieve state-of-the-art or comparable to state-of-the-art performance  ... 
arXiv:1611.06748v1 fatcat:4prwbxweyjhgdckyyndnogb5pu

A Comprehensive Review of Blind Deconvolution Techniques for Image Deblurring

Pooja Satish, Mallikarjunaswamy Srikantaswamy, Nataraj Ramaswamy
2020 Traitement du signal  
Based on the overall technique simple MAP model falls short in either deriving accurate convergence to the global optimum or computational implementation.  ...  The consequence of different priors to the rate of convergence of the MAP algorithm and its computational complexity are studied and tabulated.  ...  ACKNOWLEDGMENT The authors acknowledge with thanks JSS Academy of Technical Education, Technical Education Quality Improvement Programme (TEQIP) Cell and Visvesvaraya Technological University (VTU), Belagavi  ... 
doi:10.18280/ts.370321 fatcat:msvodhsyhrchdlpzjvvwpuffwq

PSF Recovery from Examples for Blind Super-Resolution

Isabelle Begin, Frank P. Ferrie
2007 2007 IEEE International Conference on Image Processing  
Results are compared with another method for blind super-resolution by using a series of quality measures.  ...  This paper addresses the problem of super-resolving a single image and recovering the characteristics of the sensor using a learning-based approach.  ...  The output of the algorithm is the convolution kernel, and the closest Gaussian is found by minimizing the Root-Mean-Square (RMS) distance between the Gaussian function and the values of the kernel.  ... 
doi:10.1109/icip.2007.4379855 dblp:conf/icip/BeginF07 fatcat:pkzhl7ivznbr5ivcr4jsm7744i

Fast Two-step Blind Optical Aberration Correction [article]

Thomas Eboli and Jean-Michel Morel and Gabriele Facciolo
2022 arXiv   pre-print
First, we estimate local Gaussian blur kernels for overlapping patches and sharpen them with a non-blind deblurring technique.  ...  Based on the measurements of the PSFs of dozens of lenses, these blur kernels are modeled as RGB Gaussians defined by seven parameters.  ...  Acknowledgements This work was partly financed by the DGA Astrid Maturation project "SURE-CAVI" no ANR-21-ASM3-0002, Office of Naval research grant N00014-17-1-2552.  ... 
arXiv:2208.00950v1 fatcat:prtftdis3vhyloqeanw63hs6g4

Blind De-Blurring of Microscopy Images for Cornea Cell Counting [article]

Alon Tchelet, Leonardo Mussa, Stefano Vojinovic
2020 arXiv   pre-print
Our approach is based on blind-deconvolution reconstruction that performs a depth-from-deblur so to either model Gaussian kernel or to fit kernels from an ad hoc lookup table.  ...  Unfortunately, clinical specular microscopy requires the acquisition of a large number of images at different focus depths because the curved shape of the cornea makes it impossible to acquire a single  ...  General Description of the Algorithm The method that we propose is based on blind-deconvolution with an a-priori estimation of the kernels.  ... 
arXiv:2006.03562v1 fatcat:xdjpjkporbcmdj32ntwelgpxai

Deep Super-Resolution Network for Single Image Super-Resolution with Realistic Degradations

Rao Muhammad Umer, Gian Luca Foresti, Christian Micheloni
2019 Proceedings of the 13th International Conference on Distributed Smart Cameras - ICDSC 2019  
The most of existing convolutional neural network (CNN) based SISR methods usually take an assumption that a LR image is only bicubicly down-sampled version of an HR image.  ...  To address this issue, we propose a deep SISR network that works for blur kernels of different sizes, and different noise levels in an unified residual CNN-based denoiser network, which significantly improves  ...  The architecture of UDNet [22] is consist of N residual units with 2 convolution layers each of 64 kernels by 3 × 3 filter size, and each convolution layer is preceded by the parametrized rectified linear  ... 
doi:10.1145/3349801.3349823 dblp:conf/icdsc/UmerFM19 fatcat:ldi2ejxh35ahpa5vdzwmdyrqae

Evaluation of image deblurring algorithms for real-time applications

Giuseppe Airo Farulla, Marco Indaco, Daniele Rolfo, Ludovico Orlando Russo, Pascal Trotta
2014 2014 9th IEEE International Conference on Design & Technology of Integrated Systems in Nanoscale Era (DTIS)  
Since decades, many different theories and algorithms have been proposed with the aim of retrieving latent images from blurry inputs; most of them work quite well, but very often incur in large execution  ...  the whole image processing architecture, in terms of throughput.  ...  using 2-D convolution (as in (1) ) with a motion kernel.  ... 
doi:10.1109/dtis.2014.6850668 dblp:conf/dtis/FarullaIRRT14 fatcat:htgtjzs4ajfrvjabskylfyxknm

DeblurGAN: Blind Motion Deblurring Using Conditional Adversarial Networks [article]

Orest Kupyn, Volodymyr Budzan, Mykola Mykhailych, Dmytro Mishkin, Jiri Matas
2018 arXiv   pre-print
The learning is based on a conditional GAN and the content loss . DeblurGAN achieves state-of-the art performance both in the structural similarity measure and visual appearance.  ...  The quality of the deblurring model is also evaluated in a novel way on a real-world problem -- object detection on (de-)blurred images.  ...  of Upper Austria in the frame of the COMET center, the CTU student grant SGS17/185/OHK3/3T/13.  ... 
arXiv:1711.07064v4 fatcat:nzcsivvhyjcn5jhsqswablshm4

Commutability of Blur and Affine Warping in Super-Resolution With Application to Joint Estimation of Triple-Coupled Variables

Xuesong Zhang, Jing Jiang, Silong Peng
2012 IEEE Transactions on Image Processing  
The main difficulty is that state-of-the-art deconvolution methods cannot be straightforwardly generalized to cope with the space-variant motion.  ...  This paper proposes a new approach to the image blind super-resolution (BSR) problem in the case of affine interframe motion.  ...  Lam and three reviewers for their constructive comments and suggestions that helped improve the clarity of presentation of this paper.  ... 
doi:10.1109/tip.2011.2174371 pmid:22067365 fatcat:qbsvqemobbalnevi62ulid4drq
« Previous Showing results 1 — 15 out of 276 results