no code implementations • 28 May 2024 • Lea Bogensperger, Dominik Narnhofer, Alexander Falk, Konrad Schindler, Thomas Pock
Medical image segmentation is a crucial task that relies on the ability to accurately identify and isolate regions of interest in medical images.
no code implementations • 19 Mar 2024 • Filip Ilic, He Zhao, Thomas Pock, Richard P. Wildes
Global obfuscation hides privacy sensitive regions, but also contextual regions important for action recognition.
no code implementations • 14 Nov 2023 • Robert Harb, Thomas Pock, Heimo Müller
They allow the creation of synthesized copies of datasets that can be shared without violating privacy regulations.
1 code implementation • 19 Oct 2023 • Martin Zach, Erich Kobler, Antonin Chambolle, Thomas Pock
In this work we tackle the problem of estimating the density $ f_X $ of a random variable $ X $ by successive smoothing, such that the smoothed random variable $ Y $ fulfills the diffusion partial differential equation $ (\partial_t - \Delta_1)f_Y(\,\cdot\,, t) = 0 $ with initial condition $ f_Y(\,\cdot\,, 0) = f_X $.
1 code implementation • 29 Jun 2023 • Edi Muškardin, Martin Tappler, Ingo Pill, Bernhard K. Aichernig, Thomas Pock
We examine the assumption that the hidden-state vectors of recurrent neural networks (RNNs) tend to form clusters of semantically similar vectors, which we dub the clustering hypothesis.
1 code implementation • 25 May 2023 • Tim Tsz-Kit Lau, Han Liu, Thomas Pock
We study the problem of approximate sampling from non-log-concave distributions, e. g., Gaussian mixtures, which is often challenging even in low dimensions due to their multimodality.
no code implementations • 10 Mar 2023 • Lea Bogensperger, Dominik Narnhofer, Filip Ilic, Thomas Pock
Medical image segmentation is a crucial task that relies on the ability to accurately identify and isolate regions of interest in medical images.
no code implementations • 21 Feb 2023 • Erich Kobler, Thomas Pock
In this paper, we propose a unified framework of denoising score-based models in the context of graduated non-convex energy minimization.
no code implementations • 16 Feb 2023 • Martin Zach, Thomas Pock, Erich Kobler, Antonin Chambolle
In this work we tackle the problem of estimating the density $f_X$ of a random variable $X$ by successive smoothing, such that the smoothed random variable $Y$ fulfills $(\partial_t - \Delta_1)f_Y(\,\cdot\,, t) = 0$, $f_Y(\,\cdot\,, 0) = f_X$.
no code implementations • 23 Dec 2022 • Dominik Narnhofer, Andreas Habring, Martin Holler, Thomas Pock
The proposed method employs estimates of the posterior variance together with techniques from conformal prediction in order to obtain coverage guarantees for the error bounds, without making any assumption on the underlying data distribution.
no code implementations • 25 Oct 2022 • Martin Zach, Florian Knoll, Thomas Pock
We propose a novel deep neural network based regularizer which is trained in a generative setting on reference magnitude images only.
1 code implementation • 13 Jul 2022 • Filip Ilic, Thomas Pock, Richard P. Wildes
Presently, a methodology and corresponding dataset to isolate the effects of dynamic information in video are missing.
no code implementations • 23 Mar 2022 • Martin Zach, Erich Kobler, Thomas Pock
We apply the regularizer to limited-angle and few-view CT reconstruction problems, where it outperforms traditional reconstruction algorithms by a large margin.
no code implementations • 22 Feb 2021 • Thomas Grandits, Simone Pezzuto, Francisco Sahli Costabal, Paris Perdikaris, Thomas Pock, Gernot Plank, Rolf Krause
In this work, we employ a recently developed approach, called physics informed neural networks, to learn the fiber orientations from electroanatomical maps, taking into account the physics of the electrical wave propagation.
no code implementations • 12 Feb 2021 • Dominik Narnhofer, Alexander Effland, Erich Kobler, Kerstin Hammernik, Florian Knoll, Thomas Pock
To this end, we solve the linear inverse problem of undersampled MRI reconstruction in a variational setting.
no code implementations • 29 Jan 2021 • Vladimir Kolmogorov, Thomas Pock
In case $h^*$ is the indicator function of a linear constraint and function $f$ is quadratic, we show a $O(1/n^2)$ convergence rate on the dual objective, requiring $O(n \log n)$ calls of $lmo$.
Optimization and Control
no code implementations • 12 Nov 2020 • Thomas Pinetz, Erich Kobler, Thomas Pock, Alexander Effland
We propose a novel learning-based framework for image reconstruction particularly designed for training without ground truth data, which has three major building blocks: energy-based learning, a patch-based Wasserstein loss functional, and shared prior learning.
no code implementations • 23 Oct 2020 • Christian Sormann, Patrick Knöbelreiter, Andreas Kuhn, Mattia Rossi, Thomas Pock, Friedrich Fraundorfer
We therefore show how we can calculate a normalization based on the expected 3D error, which we can then use to normalize the label jumps in the CRF.
1 code implementation • 15 Jun 2020 • Erich Kobler, Alexander Effland, Karl Kunisch, Thomas Pock
In this work, we combine the variational formulation of inverse problems with deep learning by introducing the data-driven general-purpose total deep variation regularizer.
1 code implementation • 13 Mar 2020 • Patrick Knöbelreiter, Christian Sormann, Alexander Shekhovtsov, Friedrich Fraundorfer, Thomas Pock
It has been proposed by many researchers that combining deep neural networks with graphical models can create more efficient and better regularized composite models.
1 code implementation • CVPR 2020 • Erich Kobler, Alexander Effland, Karl Kunisch, Thomas Pock
Diverse inverse problems in imaging can be cast as variational problems composed of a task-specific data fidelity term and a regularization term.
no code implementations • ECCV 2020 • Markus Hofinger, Samuel Rota Bulò, Lorenzo Porzi, Arno Knapitsch, Thomas Pock, Peter Kontschieder
In this work we review the coarse-to-fine spatial feature pyramid concept, which is used in state-of-the-art optical flow estimation networks to make exploration of the pixel flow search space computationally tractable and efficient.
no code implementations • 2 Oct 2019 • Thomas Pinetz, Daniel Soukup, Thomas Pock
Generative Adversarial Networks (GANs) have been used to model the underlying probability distribution of sample based datasets.
no code implementations • 25 Sep 2019 • Thomas Pinetz, Daniel Soukup, Thomas Pock
The goal of generative models is to model the underlying data distribution of a sample based dataset.
no code implementations • 31 Jul 2019 • Patrick Knöbelreiter, Thomas Pock
The proposed variational network arises naturally from unrolling the iterates of a proximal gradient method applied to a variational energy defined in a joint disparity, color, and confidence image space.
no code implementations • 29 Jul 2019 • Patrick Knöbelreiter, Christoph Vogel, Thomas Pock
Recent developments established deep learning as an inevitable tool to boost the performance of dense matching and stereo estimation.
no code implementations • 19 Jul 2019 • Alexander Effland, Erich Kobler, Karl Kunisch, Thomas Pock
We investigate a well-known phenomenon of variational approaches in image processing, where typically the best image quality is achieved when the gradient flow process is stopped before converging to a stationary point.
1 code implementation • NeurIPS 2019 • K. S. Sesh Kumar, Francis Bach, Thomas Pock
We consider the problem of minimizing the sum of submodular set functions assuming minimization oracles of each summand function.
2 code implementations • 6 Apr 2019 • Mahesh Chandra Mukkamala, Peter Ochs, Thomas Pock, Shoham Sabach
Backtracking line-search is an old yet powerful strategy for finding a better step sizes to be used in proximal gradient algorithms.
no code implementations • 1 Apr 2019 • Florian Knoll, Kerstin Hammernik, Chi Zhang, Steen Moeller, Thomas Pock, Daniel K. Sodickson, Mehmet Akcakaya
Both linear and non-linear methods are covered, followed by a discussion of recent efforts to further improve parallel imaging using machine learning, and specifically using artificial neural networks.
1 code implementation • 9 Nov 2018 • Christoph Vogel, Patrick Knöbelreiter, Thomas Pock
Modern optical flow methods are often composed of a cascade of many independent steps or formulated as a black box neural network that is hard to interpret and analyze.
1 code implementation • 9 Apr 2018 • Katrin Lasinger, Christoph Vogel, Thomas Pock, Konrad Schindler
We show, for the first time, how to jointly reconstruct both the individual tracer particles and a dense 3D fluid motion field from the image data, using an integrated energy minimization.
no code implementations • 9 Apr 2018 • Katrin Lasinger, Christoph Vogel, Thomas Pock, Konrad Schindler
We propose a new method for iterative particle reconstruction (IPR), in which the locations and intensities of all particles are inferred in one joint energy minimization.
no code implementations • 13 Feb 2018 • Markus Hofinger, Thomas Pock, Thomas Moosbrugger
Wood-composite materials are widely used today as they homogenize humidity related directional deformations.
no code implementations • 4 Oct 2017 • Audrey Richard, Christoph Vogel, Maros Blaha, Thomas Pock, Konrad Schindler
We propose a novel framework for the discretisation of multi-label problems on arbitrary, continuous domains.
no code implementations • ICCV 2017 • Stefan Heber, Wei Yu, Thomas Pock
In the first part the network encodes relevant information from the given input into a set of high-level feature maps.
no code implementations • 20 Jul 2017 • Gottfried Munda, Alexander Shekhovtsov, Patrick Knöbelreiter, Thomas Pock
We tackle the computation- and memory-intensive operations on the 4D cost volume by a min-projection which reduces memory complexity from quadratic to linear and binary descriptors for efficient matching.
2 code implementations • 3 Apr 2017 • Kerstin Hammernik, Teresa Klatzer, Erich Kobler, Michael P. Recht, Daniel K. Sodickson, Thomas Pock, Florian Knoll
Due to its high computational performance, i. e., reconstruction time of 193 ms on a single graphics card, and the omission of parameter tuning once the network is trained, this new approach to image reconstruction can easily be integrated into clinical workflow.
1 code implementation • 15 Mar 2017 • Christian Reinbacher, Gottfried Munda, Thomas Pock
In this work we propose a novel method to perform camera tracking of event cameras in a panoramic setting with three degrees of freedom.
2 code implementations • 8 Feb 2017 • Thomas Pock, Shoham Sabach
In this paper we study nonconvex and nonsmooth optimization problems with semi-algebraic data, where the variables vector is split into several blocks of variables.
Optimization and Control
no code implementations • CVPR 2017 • Patrick Knöbelreiter, Christian Reinbacher, Alexander Shekhovtsov, Thomas Pock
We propose a novel and principled hybrid CNN+CRF model for stereo estimation.
1 code implementation • 21 Jul 2016 • Christian Reinbacher, Gottfried Graber, Thomas Pock
In our experiments we verify that solving the variational model on the manifold produces high-quality images without explicitly estimating optical flow.
no code implementations • CVPR 2016 • Maros Blaha, Christoph Vogel, Audrey Richard, Jan D. Wegner, Thomas Pock, Konrad Schindler
We propose an adaptive multi-resolution formulation of semantic 3D reconstruction.
no code implementations • CVPR 2016 • Stefan Heber, Thomas Pock
In this paper we utilize CNNs to predict depth information for given Light Field (LF) data.
no code implementations • 23 Jan 2016 • Alexander Shekhovtsov, Christian Reinbacher, Gottfried Graber, Thomas Pock
Dense image matching is a fundamental low-level problem in Computer Vision, which has received tremendous attention from both discrete and continuous optimization communities.
no code implementations • 20 Nov 2015 • Tuomo Valkonen, Thomas Pock
We propose several variants of the primal-dual method due to Chambolle and Pock.
no code implementations • 12 Aug 2015 • Yunjin Chen, Thomas Pock
The TNRD approach is applicable for a variety of image restoration tasks by incorporating appropriate reaction force.
Ranked #6 on Color Image Denoising on Darmstadt Noise Dataset
no code implementations • CVPR 2015 • Gottfried Graber, Jonathan Balzer, Stefano Soatto, Thomas Pock
We propose a method for dense three-dimensional surface reconstruction that leverages the strengths of shape-based approaches, by imposing regularization that respects the geometry of the surface, and the strength of depth-map-based stereo, by avoiding costly computation of surface topology.
3 code implementations • CVPR 2015 • Yunjin Chen, Wei Yu, Thomas Pock
We propose to train the parameters of the filters and the influence functions through a loss based approach.
no code implementations • 26 Feb 2015 • Vladimir Kolmogorov, Thomas Pock, Michal Rolinek
We consider the problem of minimizing the continuous valued total variation subject to different unary terms on trees and propose fast direct algorithms based on dynamic programming to solve these problems.
no code implementations • 21 Apr 2014 • Yunjin Chen, Wensen Feng, René Ranftl, Hong Qiao, Thomas Pock
The Fields of Experts (FoE) image prior model, a filter-based higher-order Markov Random Fields (MRF) model, has been shown to be effective for many image restoration problems.
no code implementations • 18 Apr 2014 • Peter Ochs, Yunjin Chen, Thomas Brox, Thomas Pock
A rigorous analysis of the algorithm for the proposed class of problems yields global convergence of the function values and the arguments.
no code implementations • 14 Mar 2014 • Dirk A. Lorenz, Thomas Pock
In this paper, we propose an inertial forward backward splitting algorithm to compute a zero of the sum of two monotone operators, with one of the two operators being co-coercive.
no code implementations • 16 Jan 2014 • Yunjin Chen, Thomas Pock, René Ranftl, Horst Bischof
It is now well known that Markov random fields (MRFs) are particularly effective for modeling image priors in low-level vision.
no code implementations • 16 Jan 2014 • Yunjin Chen, Thomas Pock, Horst Bischof
We then introduce an approach to learn both analysis operator and synthesis dictionary simultaneously by using a unified framework of bi-level optimization.
no code implementations • 16 Jan 2014 • Yunjin Chen, René Ranftl, Thomas Pock
Inpainting based image compression approaches, especially linear and non-linear diffusion models, are an active research topic for lossy image compression.
no code implementations • 13 Jan 2014 • Yunjin Chen, René Ranftl, Thomas Pock
Numerical experiments show that our trained models clearly outperform existing analysis operator learning approaches and are on par with state-of-the-art image denoising algorithms.
no code implementations • CVPR 2013 • Peter Ochs, Alexey Dosovitskiy, Thomas Brox, Thomas Pock
Here we extend the problem class to linearly constrained optimization of a Lipschitz continuous function, which is the sum of a convex function and a function being concave and increasing on the non-negative orthant (possibly non-convex and nonconcave on the whole space).
no code implementations • 26 Apr 2013 • Peter Innerhofer, Thomas Pock
In this paper we propose a global convex approach for image hallucination.
no code implementations • 26 Apr 2013 • Gernot Riegler, Thomas Pock, Werner Pötzi, Astrid Veronig
The information produced by our method can be used for near real-time alerts and the statistical analysis of existing data by solar physicists.