Compressive sensing for polarization sensitive optical coherence tomography.

Area of Science:

• Biomedical Optics
• Medical Imaging Technology

Background:

• Polarization-sensitive optical coherence tomography (PS-OCT) provides valuable microstructural and birefringence information.
• Acquiring full volumetric PS-OCT datasets requires numerous B-scans, limiting imaging speed.
• Compressive sensing (CS) and sparse sampling offer potential solutions for accelerated data acquisition.

Purpose of the Study:

• To implement and evaluate compressive sensing (CS) and sparse sampling for reducing B-scan acquisition in volumetric PS-OCT.
• To assess the feasibility of reconstructing PS-OCT data with significantly fewer B-scans.
• To determine the impact of reduced B-scan rates on the accuracy and quality of reconstructed PS-OCT measurements.

Main Methods:

• Implemented sparse sampling by randomizing step sizes along the slow-axis of PS-OCT imaging.
• Applied compressive sensing (CS) algorithms to reconstruct volumetric PS-OCT data from undersampled B-scans.
• Evaluated reconstruction accuracy at missing B-scan rates of 25%, 50%, and 75% using correlation coefficients.

Main Results:

• CS successfully reconstructed reasonably good PS-OCT measurements (correlation coefficient >0.6) even with up to 50% missing B-scans.
• Sparse sampling with CS enabled a significant reduction in the number of required B-scans.
• The method demonstrated the potential to accelerate volumetric PS-OCT measurements, effectively doubling the acquisition rate.

Conclusions:

• Compressive sensing combined with sparse sampling is an effective strategy for accelerating volumetric PS-OCT acquisition.
• This approach allows for high-quality PS-OCT data reconstruction with substantially reduced scan times.
• The findings suggest a viable pathway for faster in-vivo and ex-vivo PS-OCT imaging applications.