Maximum entropy methods for texture synthesis: theory and practice
From MaRDI portal
convolutional neural networkinformation geometryLangevin algorithmMarkov chainsmaximum entropytexture synthesis
Discrete-time Markov processes on general state spaces (60J05) Markov chains (discrete-time Markov processes on discrete state spaces) (60J10) Computational methods in Markov chains (60J22) Statistical aspects of information-theoretic topics (62B10) Stochastic approximation (62L20) Applications of statistics (62P99) Monte Carlo methods (65C05) Numerical analysis or methods applied to Markov chains (65C40) Artificial neural networks and deep learning (68T07) Computing methodologies for image processing (68U10) Image processing (compression, reconstruction, etc.) in information and communication theory (94A08) Measures of information, entropy (94A17) Sampling theory in information and communication theory (94A20)
Abstract: Recent years have seen the rise of convolutional neural network techniques in exemplar-based image synthesis. These methods often rely on the minimization of some variational formulation on the image space for which the minimizers are assumed to be the solutions of the synthesis problem. In this paper we investigate, both theoretically and experimentally, another framework to deal with this problem using an alternate sampling/minimization scheme. First, we use results from information geometry to assess that our method yields a probability measure which has maximum entropy under some constraints in expectation. Then, we turn to the analysis of our method and we show, using recent results from the Markov chain literature, that its error can be explicitly bounded with constants which depend polynomially in the dimension even in the non-convex setting. This includes the case where the constraints are defined via a differentiable neural network. Finally, we present an extensive experimental study of the model, including a comparison with state-of-the-art methods and an extension to style transfer.
Recommendations
- A survey of exemplar-based texture synthesis
- Wasserstein generative models for patch-based texture synthesis
- Variational texture synthesis with sparsity and spectrum constraints
- Wasserstein loss for image synthesis and restoration
- A texture synthesis model based on semi-discrete optimal transport in patch space
Cites work
- A conditional multiscale locally Gaussian texture synthesis algorithm
- A parametric texture model based on joint statistics of complex wavelet coefficients
- A Stochastic Approximation Method
- A survey of exemplar-based texture synthesis
- A texture synthesis model based on semi-discrete optimal transport in patch space
- Convergence and robustness of the Robbins-Monro algorithm truncated at randomly varying bounds
- Efficient stochastic optimisation by unadjusted Langevin Monte Carlo. Application to maximum marginal likelihood and empirical Bayesian estimation
- Equivalence of Julesz ensembles and FRAME models
- Exponential convergence in L^p-Wasserstein distance for diffusion processes without uniformly dissipative drift
- Exponential convergence of Langevin distributions and their discrete approximations
- Fast Langevin based algorithm for MCMC in high dimensions
- scientific article; zbMATH DE number 3621626 (Why is no real title available?)
- scientific article; zbMATH DE number 1114382 (Why is no real title available?)
- scientific article; zbMATH DE number 1153603 (Why is no real title available?)
- scientific article; zbMATH DE number 1560711 (Why is no real title available?)
- scientific article; zbMATH DE number 3800753 (Why is no real title available?)
- I-divergence geometry of probability distributions and minimization problems
- Information Theory and Statistical Mechanics
- Introductory lectures on convex optimization. A basic course.
- Markov Chains
- Maximum likelihood estimation of regularization parameters in high-dimensional inverse problems: an empirical Bayesian approach. II: Theoretical analysis
- MCMC using Hamiltonian dynamics
- Monte Carlo sampling methods using Markov chains and their applications
- Multiscale sparse microcanonical models
- Nonasymptotic bounds for sampling algorithms without log-concavity
- Nonasymptotic convergence analysis for the unadjusted Langevin algorithm
- On the Existence and Characterization of the Maxent Distribution Under General Moment Inequality Constraints
- Pattern theory. The stochastic analysis of real-world signals.
- Problems and theorems in analysis. I. Series, integral calculus, theory of functions. Transl. from the German by Dorothee Aeppli
- Quantitative contraction rates for Markov chains on general state spaces
- Random Phase Textures: Theory and Synthesis
- Reflection coupling and Wasserstein contractivity without convexity
- Reflection couplings and contraction rates for diffusions
- Riemann manifold Langevin and Hamiltonian Monte Carlo methods. With discussion and authors' reply
- Sanov property, generalized I-projection and a conditional limit theorem
- Texture synthesis and nonparametric resampling of random fields
- The geometric foundations of Hamiltonian Monte Carlo
- Theoretical Guarantees for Approximate Sampling from Smooth and Log-Concave Densities
Cited in
(6)- Constrained image generation using binarized neural networks with decision procedures
- Multimodal image synthesis with conditional implicit maximum likelihood estimation
- High-resolution neural texture synthesis with long-range constraints
- Wasserstein loss for image synthesis and restoration
- ON THE SOLVABILITY OF MAXIMUM ENTROPY MOMENT PROBLEMS IN TEXTURE ANALYSIS
- Macrocanonical models for texture synthesis
This page was built for publication: Maximum entropy methods for texture synthesis: theory and practice
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4999346)
