Automated extraction of 3D Doppler OCT signatures using a support vector machine

摘要

We present a method for the automated extraction of Doppler OCT flow information by using a support vector machine that combines different features for classification. We employ histogram equalization that makes it possible to distinguish vessels from bulk tissue by texture analysis. This method is particularly applicable to settings with significant phase noise as it is more robust to multiple scattering components than simple threshold-based methods.