Phase-resolved Doppler optical coherence tomography--limitations and improvements.

We describe how the simple phase difference averaging causes a systematic bias in the velocity estimation obtained by phase-resolved Fourier domain optical coherence tomography (FdOCT). The magnitude of this bias depends on the signal-to-noise ratio as well as proximity of the measured velocity to the limits of the velocity range. We demonstrate the proper way of data processing, which enables obtaining velocity values free of this error. We validate the improved technique by measurements of flow velocity in glass capillaries, in human retinal vessels, and we compare the results with those obtained by standard phase-resolved FdOCT.