A smart diagnostic system for Doppler ultrasonography

This paper proposes an automated diagnostic approach for a Doppler ultrasonography system (DUS) using the spectrogram features and image information. Various boundary and moment features of the spectrogram are fed to a supervised ANN to diagonise the localized arterial condition. This local data is given to a Bayesian probabilistic inferencing engine for end diagnosis. Simultaneously, arterial contour variation in the image is tracked for motion estimation and diagnosis of the arterial condition. Though the exact correlation remain unexplored, but the two findings amply help in medical diagnosis of arterial ailments