An optimization method with precomputed starting points for solving the inverse Mie problem

Abstract

We study the inverse light-scattering problem which arises in the characterization of small particles by means of scanning flow cytometry. The problem is stated in general form as the problem of solution of a nonlinear equation and solved by the gradient optimization method. In this event, the problem of choosing the starting point appears. In this paper, we propose a method for making this choice based on the preliminary analysis of the direct map. A number of numerical examples are given, using both synthetic and experimental data.

Single‐Particle Characterization by Elastic Light Scattering

“…Another issue is the choice of the initial guess, which can accelerate global optimization and is critical for local‐optimization algorithms. This is commonly solved based on the prior information, but sometimes using another (approximate) characterization method, for instance, a spectral sizing, [ 202 ] the fitting by a simpler model (sphere), [ 53 ] the nearest‐neighbor interpolation using a small dataset, [ 203 ] or image processing tools. [ 137 ] Let us further take a closer look at various examples of regression techniques.…”