GPU-based computational adaptive optics for volumetric optical coherence microscopy

Han Tang; Jeffrey A. Mulligan; Gavrielle R. Untracht; Xihao Zhang; Steven G. Adie

doi:10.1117/12.2213949

7 March 2016 GPU-based computational adaptive optics for volumetric optical coherence microscopy

Han Tang, Jeffrey A. Mulligan, Gavrielle R. Untracht, Xihao Zhang, Steven G. Adie

Proceedings Volume 9720, High-Speed Biomedical Imaging and Spectroscopy: Toward Big Data Instrumentation and Management; 97200O (2016) https://doi.org/10.1117/12.2213949
Event: SPIE BiOS, 2016, San Francisco, California, United States

Abstract

Optical coherence tomography (OCT) is a non-invasive imaging technique that measures reflectance from within biological tissues. Current higher-NA optical coherence microscopy (OCM) technologies with near cellular resolution have limitations on volumetric imaging capabilities due to the trade-offs between resolution vs. depth-of-field and sensitivity to aberrations. Such trade-offs can be addressed using computational adaptive optics (CAO), which corrects aberration computationally for all depths based on the complex optical field measured by OCT. However, due to the large size of datasets plus the computational complexity of CAO and OCT algorithms, it is a challenge to achieve high-resolution 3D-OCM reconstructions at speeds suitable for clinical and research OCM imaging. In recent years, real-time OCT reconstruction incorporating both dispersion and defocus correction has been achieved through parallel computing on graphics processing units (GPUs). We add to these methods by implementing depth-dependent aberration correction for volumetric OCM using plane-by-plane phase deconvolution. Following both defocus and aberration correction, our reconstruction algorithm achieved depth-independent transverse resolution of 2.8 um, equal to the diffraction-limited focal plane resolution. We have translated the CAO algorithm to a CUDA code implementation and tested the speed of the software in real-time using two GPUs - NVIDIA Quadro K600 and Geforce TITAN Z. For a data volume containing 4096×256×256 voxels, our system’s processing speed can keep up with the 60 kHz acquisition rate of the line-scan camera, and takes 1.09 seconds to simultaneously update the CAO correction for 3 en face planes at user-selectable depths.

Conference Presentation

Citation Download Citation

Han Tang, Jeffrey A. Mulligan, Gavrielle R. Untracht, Xihao Zhang, and Steven G. Adie "GPU-based computational adaptive optics for volumetric optical coherence microscopy", Proc. SPIE 9720, High-Speed Biomedical Imaging and Spectroscopy: Toward Big Data Instrumentation and Management, 97200O (7 March 2016); https://doi.org/10.1117/12.2213949

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available