Existing models of image formation in optical coherence tomography are based upon the extended Huygens-Fresnel formalism. These models all, to varying degrees, rely on scatterer ensemble averages, rather than deterministic scattering distributions. Whilst the former is sometimes preferable, there are a growing number of applications where the ability to predict image formation based upon deterministic refractive index distributions is of great interest, including, for example, image formation in turbid tissue.
A rigorous model based upon three-dimensional solutions of Maxwell's equations offers a number of tantalising opportunities. For example, shedding light on features near or below the resolution of an OCT system and on the impact of phenomena usually described as diffraction, interference and scattering, but which more generally result from light scattering satisfying Maxwell's equations. A rigorous model allows inverse scattering methods to be developed not requiring the first-order Born approximation. Finally, a rigorous model can provide gold standard verification of myriad quantitative techniques currently being developed throughout the field.
We have developed the first such model of image formation based upon three-dimensional solutions of Maxwell's equations, which has vastly different properties to models based on two-dimensional solutions. Although we present simulated B-scans, this model is equally applicable to C-scans. This has been made possible by advances in computational techniques and in computational resources routinely available. We will present the main features of our model, comparisons of measured and simulated image formation for phantoms and discuss the future of rigorous modelling in optical coherence tomography research and application.
|