ニューラルネットワークによる超音波像組織性状識別法の検討

抄録

We have classified parenchymal echo patterns of cirrhotic liver into four types according to the size of hypoechoic nodular lesions. Neural network technique has been applied to the characterization of hepatic parenchymal diseases in ultrasonic B-scan textures. We employed a multilayer feedforward neural network utilizing the back-propagation algorithm. Coarse score of cirrhotic liver at autopsy correlates closely with histological findings. The coarse score is also bound up with parenchymal echo pattern of the liver. The experimental study suggests that the neural network approach is useful for objective evaluation of hepatic parenchymal echo pattern, though continuing work is needed to improve the classification accuracy.