超音波映像セルラニューラルネットワークを利用いたスペックルノイズ低減に関する研究(医・工学連携における新しい計測・解析技術開発への期待)

概要

論文の詳細を見る
This paper is for speckle noise reduction and emphasis of boundary in medical ultrasound images. Generally, in case of speckle reduction, it often happened to lose the boundary of organization or small tumor because of blurring from noise reduction. Then, in this paper, we proposed cellular neural network based new method to reduce speckle of ultrasound image with strengthening the boundary. The proposed method is consists of supervised learning to decide two templates of cellular neural network with back propagation and iterative execution of cellular neural network. To make supervised learning of cellular neural network, we used 9 patterns images. And we took five times iteration for experimentation. From the comparison of results with existing previous methods such as PERONA and SRAD, we could have verified the worth of proposed method.
2007-01-19