Learning Gaussian Conditional Random Fields for Low-Level Vision

By Marshall F. Tappen, Ce Liu, Edward H. Adelson, William T. Freeman. 2007

170 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Meta-Learning With Differentiable Convex Optimization (2019, methodology): “This has been applied for learning in low-level vision problems [30, 27].”
OptNet: Differentiable Optimization as a Layer in Neural Networks (2017, background): “…Brakel et al., 2013; LeCun et al., 2006; Belanger & McCallum, 2016; Belanger et al., 2017; Amos et al., 2017); in vision for tasks such as denoising (Tappen et al., 2007; Schmidt & Roth, 2014); and Metz et al. (2016) uses unrolled optimization within a network to stabilize the convergence of…”
Fast, Exact and Multi-scale Inference for Semantic Image Segmentation with Deep Gaussian CRFs (2016, background): “Motivated by [16,17], our starting point in this work is the observation that a particular type of graphical model, the Gaussian Conditional Random Field (G-CRF), allows us to perform exact and efficient Maximum-A-Posteriori (MAP) inference.”
Natural Image Denoising with Convolutional Networks (2008, methodology): “Initial work used handdesigned settings of the parameters, but recently there has been increasing success in learning the parameters of such models from databases of natural images [3] [4] [5] [6] [7] [8].”
The structure of optimal parameters for image restoration problems (2015, background): “In [41, 42, 27, 28, 18, 19], for instance the authors consider bilevel optimization for finite dimensional Markov random field (MRF) models.”
Training Deep Networks with Structured Layers by Matrix Backpropagation (2015, background): “…case, and focusing on the subset of the elements (variables) of interest, appeared in the context of camera calibration[53], for learning parameters in a normalized cuts model[14], learning the parameters of Gaussian CRFs for denoising [43] and learning deep canonical correlation models [54].”
End-to-End Learning for Structured Prediction Energy Networks (2017, methodology): “We also can use the local terms to provide an implementation of Init ( · ) .”
Modeling Multiscale Subbands of Photographic Images with Fields of Gaussian Scale Mixtures (2009, methodology): “A number of authors have developed MRF-based image modelsinthepixeldomain(e.g.,[21],[22],[23],[24],[ 25 ],also see[26]foranoverview).Inparticular,therecentlydeveloped field of experts (FoE) model [27] has been used to achieve impressive performance in denoising.”
A Bilevel Optimization Approach for Parameter Learning in Variational Models (2013, background): “in [29, 28, 27] to learn the parameters of different Markov random field models.”
Learning Large-Scale Automatic Image Colorization (2015, background): “Learning to Optimize: A Gaussian random field (GRF) models the log-likelihood of a colorization (or other set of intrinsic image layers) c as − (1/2)c T Σ −1 (I)c − b T (I)c + K, where K is a constant of no interest; the first application to intrinsic images is by Tappen et al [18].”
Noise2Void - Learning Denoising From Single Noisy Images (2018, background): “[16, 19]) explicitly modeled these interdependencies via Markov Random Fields (MRFs).”
Beyond pixels: exploring new representations and applications for motion analysis (2009, methodology): “In Appendix B, we show that IRLS is also equivalent to variational optimization [115].”
Dense and Low-Rank Gaussian CRFs Using Deep Embeddings (2017, unknown): “be found in [2, 16, 32, 33].”
Learning optimized MAP estimates in continuously-valued MRF models (2009, methodology): “Previous authors have pointed out two important points regarding Equation 6 [3, 15].”
Gaussian Conditional Random Field Network for Semantic Segmentation (2016, methodology): “Compared to discrete CRFs, GCRFs are simpler models that can be solved optimally.”
Shrinkage Fields for Effective Image Restoration (2014, background): “Gaussian CRFs and their associated loss-based training have ﬁrst been introduced by Tappen [27].”
Who killed the directed model? (2008, unknown): “Given the above difficulties, work on learning undirected models either restricts attention to simple (e.g. Gaussian ) functions [18], uses exhaustive Markov chain Monte Carlo techniques [13] [23], variational methods [19] [22], the pseudo-likelihood approximation [5], and/or other approximations…”
Power Watershed: A Unifying Graph-Based Optimization Framework (2011, background): “These kind of MRFs have been applied in image segmentation and other pattern recognition applications [48], [5], [83].”