Diffraction tomography from intensity measurements: an evolutionary stochastic search to invert experimental data.

Area of Science:

• Computational imaging
• Wave physics
• Tomographic reconstruction

Background:

• Iterative diffraction tomography algorithms are crucial for imaging.
• Existing methods like distorted Born algorithms face challenges with noisy data and poor measurement sensitivity.

Purpose of the Study:

• To develop a robust algorithm for refractive index distribution reconstruction from intensity transport data.
• To address limitations of deterministic algorithms in handling noisy data and poor sensitivity.

Main Methods:

• Development of iterative diffraction tomography algorithms based on the distorted Born approximation.
• Implementation of a derivative-free evolutionary stochastic scheme with additive updates.
• Comparison of the stochastic algorithm against the deterministic Gauss-Newton (GN) scheme.

Main Results:

• The Born approximation provides stable forward equations and nearly oscillation-free solutions.
• The stochastic algorithm effectively bridges measurement-prediction misfit using additive updates.
• Superiority of the stochastic algorithm demonstrated through successful refractive index profile reconstruction from simulated and experimental data.

Conclusions:

• The developed stochastic algorithm offers a superior approach for refractive index imaging compared to traditional GN methods.
• This method is effective even with inherently noisy intensity transport data.
• The findings have implications for advanced imaging techniques in various scientific fields.