Resonant Doppler Fourier domain optical coherence tomography (FDOCT) is a functional imaging tool for extracting tissue flow. The method is based on the effect of interference fringe blurring in spectrometer-based FDOCT, where the path difference between structure and reference changes during camera integration. If the reference path length is changed in resonance with the Doppler frequency of the sample flow, the signals of resting structures will be suppressed, whereas the signals of blood flow are enhanced. This allows for an easy extraction of vascularization structure. Conventional flow velocity analysis extracts only the axial flow component, which strongly depends on the orientation of the vessel with respect to the incident light. We introduce an algorithm to extract the vessel geometry within the 3-D data volume. The algorithm calculates the angular correction according to the local gradients of the vessel orientations. We apply the algorithm on a measured 3-D resonant Doppler dataset. For validation of the reproducibility, we compare two independently obtained 3-D flow maps of the same volunteer and region.