Diffuse optical tomography by simulated annealing via a spin Hamiltonian.

Area of Science:

• Biomedical optics
• Medical imaging
• Computational physics

Background:

• Diffuse optical tomography (DOT) is an imaging technique utilizing near-infrared light to reconstruct optical properties of biological tissues.
• Traditional iterative methods for solving the DOT inverse problem often require accurate initial guesses of optical parameters, limiting their applicability.
• The lack of good initial guesses can lead to inaccurate or failed reconstructions in DOT.

Purpose of the Study:

• To develop a novel numerical scheme for diffuse optical tomography (DOT) that does not require initial guesses of optical parameters.
• To address the limitations of existing iterative methods in DOT reconstruction.
• To enable accurate DOT imaging even with poor or unavailable initial parameter estimates.

Main Methods:

• The study proposes a simulated annealing (SA) based numerical scheme, a type of Markov-chain Monte Carlo method.
• A spin Hamiltonian is incorporated into the cost function for SA implementation in DOT.
• The Metropolis algorithm or single-component Metropolis-Hastings algorithm is employed within the SA framework.

Main Results:

• Numerical experiments demonstrate that the SA method successfully converges from random initial spin configurations.
• The proposed method effectively reconstructs targets within the simulated medium.
• The inverse problem in DOT is solved by finding the ground state of the spin Hamiltonian using SA.

Conclusions:

• The developed SA-based numerical scheme provides a robust solution for the DOT inverse problem.
• This method overcomes the critical dependency on initial guesses inherent in conventional DOT algorithms.
• The findings suggest a promising new approach for accurate and reliable DOT imaging in various applications.