Towards the development of an automated blood vessel classification system

The implementation and optimization of a deterministic pattern recognition system to automatically classify venous and arterial signals is presented. CWT and STFT spectral analysis methods were applied to a pool of ninety-seven Doppler signals consisting of both arterial and venous signals. Subsequently, a classification algorithm was developed to examine each resulting spectral data to differentiate between the signal types. The CWT method proved more accurate than the STFT method because of its exceptional spectral resolution. The results of the study can be employed to medical systems in which distinguishing between venous and arterial signals is of primary importance.