Abstract
Current optical coherence tomography (OCT) based micro-angiography is prone to noise that arises from static background. This work presents a novel feature space based optical micro-angiography (OMAG) method (fsOMAG) that can effectively differentiate flow signal from static background in the feature space. fsOMAG consists of two steps. In the first step a classification map is generated that provides criterion for classification in the second step to extract functional blood flow from experimental data set. The performance of fsOMAG is examined through phantom experiments and in-vivo human retinal imaging, and compared with the existing OMAG. The results indicate its potential for clinical applications.