画像処理を用いた生物細胞の認識 : 第1報, ニューラル・ネットワークを用いた植物細胞の生死認識

概要

論文の詳細を見る
This paper deals with an automatic culture monitoring system of the plant cell "protoplast". The conventional monitoring system mainly has two problems : [i] no method of recognizing the life or death of a cell without using chemicals, [ii] the hard and laborious work for biologists to monitor cells and to find the best culture condition for them. In this paper, a new method using image processing and the neural network is proposed to solve these problems. The image processing can provide a method of measuring of the shape and the colors of a cell. The neural network can recognize the life and death of a cell from image measurements. This method gives correct answering at an accuracy rate of 85 percent. Because this accuracy is considered about equal to human performance, this system will provide an alternative method to monitoring cells.
一般社団法人日本機械学会の論文
1991-10-25