Correcting spatial-spectral crosstalk and chromatic aberrations in broadband line-scan spectral-domain OCT images.
Le HanKostadinka BizhevaPublished in: Biomedical optics express (2023)
Digital correction of optical aberrations allows for high-resolution imaging across the full depth range in optical coherence tomography (OCT). Many digital aberration correction (DAC) methods have been proposed in the past to evaluate and correct monochromatic error in OCT images. However, other factors that deteriorate the image quality have not been fully investigated. Specifically, in a broadband line-scan spectral-domain OCT system (LS-SD-OCT), photons with different wavelengths scattered from the same transverse location and in the imaged object will be projected onto different spatial coordinates onto the 2D camera sensor, which in this work is defined as spatial-spectral crosstalk. In addition, chromatic aberrations in both axial and lateral directions are not negligible for broad spectral bandwidths. Here we present a novel approach to digital recovery of the spatial resolution in images acquired with a broadband LS-SD-OCT, which addresses these two main factors that limit the effectiveness of DAC for restoring diffraction-limited resolution in LS-SD-OCT images. In the proposed approach, spatial-spectral crosstalk and chromatic aberrations are suppressed by the registration of monochromatic sub-band tomograms that are digitally corrected for aberrations. The new method was validated by imaging a standard resolution target, a microspheres phantom, and different biological tissues. LS-SD-OCT technology combined with the proposed novel image reconstruction method could be a valuable research tool for various biomedical and clinical applications.
Keyphrases
- optical coherence tomography
- high resolution
- diabetic retinopathy
- image quality
- optic nerve
- computed tomography
- copy number
- high speed
- dual energy
- randomized controlled trial
- deep learning
- single molecule
- systematic review
- gene expression
- climate change
- minimally invasive
- genome wide
- mass spectrometry
- machine learning
- magnetic resonance
- convolutional neural network
- dna methylation