Paper Info
Title
Super-Resolution Reconstruction in Frequency-Domain Optical-Coherence Tomography Using the Finite-Rate-of-Innovation Principle
Abstract
The standard approach to signal reconstruction in frequency-domain optical-coherence tomography (FDOCT) is to apply the inverse Fourier transform to the measurements. This technique offers limited resolution (due to Heisenberg's uncertainty principle). We propose a new super-resolution reconstruction method based on a parametric representation. We consider multilayer specimens, wherein each layer has a constant refractive index and show that the backscattered signal from such a specimen fits accurately in to the framework of finite-rate-of-innovation (FRI) signal model and is represented by a finite number of free parameters. We deploy the high-resolution Prony method and show that high-quality, super-resolved reconstruction is possible with fewer measurements (about one-fourth of the number required for the standard Fourier technique). To further improve robustness to noise in practical scenarios, we take advantage of an iterated singular-value decomposition algorithm (Cadzow denoiser). We present results of Monte Carlo analyses, and assess statistical efficiency of the reconstruction techniques by comparing their performance against the Cramér-Rao bound. Reconstruction results on experimental data obtained from technical as well as biological specimens show a distinct improvement in resolution and signal-to-reconstruction noise offered by the proposed method in comparison with the standard approach.
Year
DOI
Venue
2014
10.1109/TSP.2014.2340811
Signal Processing, IEEE Transactions  
Keywords
DocType
Volume
Monte Carlo methods,frequency-domain analysis,iterative methods,medical signal processing,optical tomography,refractive index,signal reconstruction,signal representation,signal resolution,singular value decomposition,Cadzow denoiser,Cramer-Rao bound,Monte Carlo analysis,backscattered signal,constant refractive index,finite-rate-of-innovation principle,finite-rate-of-innovation signal model,frequency-domain optical-coherence tomography,high-resolution Prony method,inverse Fourier transform,iterated singular-value decomposition algorithm,parametric representation,signal reconstruction,signal resolution,super-resolution reconstruction,Annihilating filter,Cadzow denoising,finite rate of innovation (FRI),frequency-domain optical-coherence tomography (FDOCT),high-resolution spectral estimation
Journal
62
Issue
ISSN
Citations 
19
1053-587X
0
PageRank 
References 
Authors
0.34
0
2
Name
Order
Citations
PageRank
Chandra Sekhar Seelamantula1487.95
Satish Mulleti2122.99