Motion correction in optical coherence tomography volumes on a per A-scan basis using orthogonal scan patterns

Area of Science:

• Ophthalmology
• Medical Imaging
• Biomedical Engineering

Background:

• High-speed Optical Coherence Tomography (OCT) enables rapid acquisition of 3D volumetric retinal data.
• Eye movement during OCT scans causes distortions, limiting the accuracy of quantitative 3D data analysis.
• Existing methods struggle to fully correct motion artifacts in densely sampled volumetric OCT datasets.

Purpose of the Study:

• To develop and evaluate a novel software-based method for retrospective motion artifact correction in 3D OCT raster scans.
• To improve the accuracy of quantitative measurements from in vivo volumetric OCT data.
• To enhance the signal quality of 3D OCT volumes by merging corrected datasets.

Main Methods:

• Retrospective motion compensation using image registration algorithms applied directly to OCT data.
• Registration of multiple, successively acquired 3D OCT volumes with orthogonal fast scan directions to estimate eye motion.
• Optimization of a global objective function with dense displacement fields and time-structured regularization for motion correction.
• Merging of motion-corrected volumes to create a single, high-quality dataset.

Main Results:

• The algorithm successfully corrects motion artifacts in all three dimensions on a per A-scan basis.
• Corrected OCT volumes exhibit no visible motion artifacts.
• Merging multiple motion-corrected volumes results in superior signal quality compared to individual input volumes.
• Demonstrated effectiveness on both spectral and swept-source OCT instruments across different retinal regions.

Conclusions:

• The developed software method effectively corrects motion artifacts in 3D OCT data.
• Motion correction and merging significantly improve image quality and should enhance morphometric measurement accuracy.
• This approach offers a valuable tool for more reliable quantitative analysis of volumetric OCT imaging.