Bayesian approach to analyzing holograms of colloidal particles

Area of Science:

• Physics, Optics
• Materials Science
• Computational Science

Background:

• Digital holography is a powerful technique for imaging microscopic objects.
• Accurate tracking and characterization of colloidal particles are essential in various scientific fields.
• Traditional methods often require precise initial parameter estimates and struggle with uncertainty quantification.

Purpose of the Study:

• To develop a robust Bayesian framework for analyzing colloidal particles from in-line digital holograms.
• To improve the accuracy of particle position, size, and refractive index determination.
• To provide a more reliable method for uncertainty quantification in holographic particle analysis.

Main Methods:

• Modeling hologram formation using Lorenz-Mie theory.
• Employing a tempered Markov-chain Monte Carlo (MCMC) method for parameter estimation.
• Sampling posterior probability distributions for particle properties.

Main Results:

• The Bayesian approach successfully tracks and characterizes colloidal particles.
• It allows for straightforward incorporation of prior knowledge.
• Provides more accurate uncertainty estimates compared to least-squares fitting.
• Eliminates the need for accurate initial parameter guesses, simplifying the process.

Conclusions:

• The developed Bayesian method offers a significant advancement in holographic particle analysis.
• It enhances the reliability and accuracy of colloidal particle tracking and characterization.
• This approach is particularly beneficial for experiments demanding precise uncertainty quantification.