Real-time en-face Gabor optical coherence tomographic angiography on human skin using CUDA GPU.

摘要: We recently proposed an optical coherence tomographic angiography (OCTA) algorithm, Gabor (GOCTA), which can extract microvascular signals from a spectral domain directly with lower computational complexity compared to other algorithms. In this manuscript, we combine programmable swept source, OCT complex signal detecting unit, and graphics process units (GPU) achieve real-time en-face GOCTA system for human skin imaging. The source balance the A-scan rate tuning range; polarization-modulation based unit double imaging depth range, GPU accelerate data processing. C++ CUDA are used as programming platform where five parallel threads created galvo-driving generation, acquisition, transfer, processing, image display, respectively. Two queues (for raw images, respectively) improve exchange efficiency among different devices. study, acquisition time processing each 3D volume (256×304×608 pixels,) 405.3 173.7 milliseconds To best of our knowledge, is first show images covering 3×3 mm2 at refresh 2.5 Hz.