Kalman-Based Real-Time Functional Decomposition for the Spectral Calibration in Swept Source Optical Coherence Tomography.

This paper presents a real-time functional decomposition adaptive algorithm for the optimal sampling of the interferometric signal in Swept-Source Optical Coherence Tomography imaging systems, which completely eliminates the input signal dependent nonlinearities that are problematic in current state-of-the-art OCT realizations that use interpolation and resampling. The proposed adaptive calibration algorithm uses the Kalman approach to estimate the wavenumber index parameter k from the Mach-Zender Interferometer signal which is then applied to an adaptive level crossing sampler to generate a sampling clock that k-linearizes the data on real-time during the sampling process. Such a system implements an artifact-free realization of the technology removing the need for classical interpolation and resampling. The new real-time linearization scheme has the additional capability of increasing the imaging acquisition speed by 10X while providing robustness to noise, properties that are demonstrated through mathematical analysis and simulation results throughout the paper.