Maximization of signal-to-noise ratio in optical coherence tomography using a depth-dependent matched filter

We discuss and demonstrate the dependence of noise on the signal in time-domain optical coherence tomography (TDOCT). We then derive a depth-dependent matched filter to maximize the signal-to-noise ratio at every pixel in a depth scan (A-scan). We use an empirical estimate of the second order statistics of the noise in OCT images of vascular tissue to implement a depth-dependent filter that is matched to these images. The application of our filter results in an average increase of signal-to-noise ratio of about 7 dB compared to a simple averaging operation. Our filter is not specific to time-domain OCT, but it is applicable to other types of OCT systems.