An Automated Ultrasound Discrimination System of Tissue under an Obstacle by Fuzzy Reasoning
Takashi Shimizu, Kouki Nagamune, Syoji Kobashi, Katsuya Kondo, Yutaka Hata, Yuri T. Kitamura*, and Toshio Yanagida*
Graduate School of Medicine, Osaka University
Proc. of 1st Int. Conf. on Soft Computing and Intelligent Systems, CD-ROM, (2002)
This paper proposes an automated discrimination system of tissue under an obstacle from ultrasound. A present ultrasonic diagnosis system is widely used for a belly and a chest. The transcranial diagnosis system with placement free using ultrasound has not been developed. This paper describes the discrimination system of tissues under an artificial bone as preliminary experiment. The artificial bone is regarded as human skull. In the first step, a fuzzy expert system is automatically constructed without an artificial bone using four attributes: an amplitude and three frequencies. In the second step, we apply the system to tissue discrimination with an artificial bone and discriminate the unknown material. In our fuzzy rule, min operation, product operation and average operation are used. The system was applied to 240 data (six pairs with 40 trial). Consequently, the system using average operation could discriminate tissue under an artificial bone with the accuracy of 70.4%.