Imaging conductivity from current density magnitude using neural networks
From MaRDI portal
Publication:5081798
Abstract: Conductivity imaging represents one of the most important tasks in medical imaging. In this work we develop a neural network based reconstruction technique for imaging the conductivity from the magnitude of the internal current density. It is achieved by formulating the problem as a relaxed weighted least-gradient problem, and then approximating its minimizer by standard fully connected feedforward neural networks. We derive bounds on two components of the generalization error, i.e., approximation error and statistical error, explicitly in terms of properties of the neural networks (e.g., depth, total number of parameters, and the bound of the network parameters). We illustrate the performance and distinct features of the approach on several numerical experiments. Numerically, it is observed that the approach enjoys remarkable robustness with respect to the presence of data noise.
Recommendations
- A regularized weighted least gradient problem for conductivity imaging
- Construct Deep Neural Networks based on Direct Sampling Methods for Solving Electrical Impedance Tomography
- Numerical conductivity reconstruction from partial interior current density information in three dimensions
- Imaging of anisotropic conductivities from current densities in two dimensions
- Imaging anisotropic conductivities from current densities
Cites work
- 10.1162/153244303321897690
- A convergent algorithm for the hybrid problem of reconstructing conductivity from minimal interior data
- A Limited Memory Algorithm for Bound Constrained Optimization
- A regularized weighted least gradient problem for conductivity imaging
- A Stochastic Approximation Method
- A weighted minimum gradient problem with complete electrode model boundary conditions for conductivity imaging
- Acousto-electric tomography with total variation regularization
- Algorithms for solving high dimensional PDEs: from nonlinear Monte Carlo to machine learning
- An introduction to mathematics of emerging biomedical imaging
- Analysis of the generalization error: empirical risk minimization over deep artificial neural networks overcomes the curse of dimensionality in the numerical approximation of Black-Scholes partial differential equations
- Anisotropic Cheeger sets and applications
- Approximation rates for neural networks with encodable weights in smoothness spaces
- Bounding the expectation of the supremum of empirical processes indexed by Hölder classes
- Conductivity imaging with a single measurement of boundary and interior data
- Construct Deep Neural Networks based on Direct Sampling Methods for Solving Electrical Impedance Tomography
- Convergence rate analysis for deep Ritz method
- Current density impedance imaging of an anisotropic conductivity in a known conformal class
- DGM: a deep learning algorithm for solving partial differential equations
- Electrical impedance tomography
- Existence and uniqueness of minimizers of general least gradient problems
- Finite neuron method and convergence analysis
- scientific article; zbMATH DE number 1950575 (Why is no real title available?)
- scientific article; zbMATH DE number 6982909 (Why is no real title available?)
- scientific article; zbMATH DE number 1420699 (Why is no real title available?)
- scientific article; zbMATH DE number 7064043 (Why is no real title available?)
- Hybrid inverse problems and internal functionals
- Hybrid tomography for conductivity imaging
- Imaging anisotropic conductivities from current densities
- Inverse problems. Tikhonov theory and algorithms
- Measure theory and fine properties of functions
- Neural Network Learning
- Neural‐network‐based approximations for solving partial differential equations
- Numerical conductivity reconstruction from partial interior current density information in three dimensions
- Numerical solution of inverse problems by weak adversarial networks
- On a Nonlinear Partial Differential Equation Arising in Magnetic Resonance Electrical Impedance Tomography
- On the mathematical foundations of learning
- Optimization methods for large-scale machine learning
- Physics constrained learning for data-driven inverse modeling from sparse observations
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Recovering the conductivity from a single measurement of interior data
- Solving inverse-PDE problems with physics-aware neural networks
- Stability in conductivity imaging from partial measurements of one interior current
- Stability of coupled-physics inverse problems with one internal measurement
- Stability of current density impedance imaging
- Stable reconstruction of regular 1-harmonic maps with a given trace at the boundary
- Strong Solutions for PDE-Based Tomography by Unsupervised Learning
- Structural stability in a minimization problem and applications to conductivity imaging
- SwitchNet: a neural Network Model for forward and inverse scattering problems
- The anisotropic total variation flow
- The deep Ritz method: a deep learning-based numerical algorithm for solving variational problems
- The Euler-Lagrange equation for the anisotropic least gradient problem
- The Finite Element Method with Penalty
- The sizes of compact subsets of Hilbert space and continuity of Gaussian processes
- Understanding machine learning. From theory to algorithms
- Uniqueness of minimizers of weighted least gradient problems arising in hybrid inverse problems
Cited in
(15)- Divide-and-conquer DNN approach for the inverse point source problem using a few single frequency measurements
- On the recovery of internal source for an elliptic system by neural network approximation
- Solving Elliptic Problems with Singular Sources Using Singularity Splitting Deep Ritz Method
- Conductivity Imaging from Internal Measurements with Mixed Least-Squares Deep Neural Networks
- Solving Poisson problems in polygonal domains with singularity enriched physics informed neural networks
- Recovering the source term in elliptic equation via deep learning: method and convergence analysis
- Current density impedance imaging with PINNs
- On the simultaneous recovery of boundary impedance and internal conductivity
- A two-stage algorithm based on convolution and attention hybrid architecture for multi-obstacle inverse scattering problems: integrating obstacle perception and shape recognition
- Potential identification via Tikhonov-PINNs
- Solving elliptic optimal control problems via neural networks and optimality system
- Potential coefficient identification problem in parabolic equation with deep neural networks
- An iterative deep Ritz method for monotone elliptic problems
- Hybrid neural-network FEM approximation of diffusion coefficient in elliptic and parabolic problems
- Imaging anisotropic conductivity from internal measurements with mixed least-squares deep neural networks
This page was built for publication: Imaging conductivity from current density magnitude using neural networks
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5081798)
