クロレラ自動計数方式におけるニューラルネットワーク学習

URI http://repository.tsuyama-ct.ac.jp/metadata/140
FulltextURL
Title
クロレラ自動計数方式におけるニューラルネットワーク学習
Title Alternative
Learning of Neural Network on Automatic Counting of Chlorella
Author
著者 薮木 登
著者(別表記) Yabuki Noboru
著者 三木 成彦
著者(別表記) Miki Shigehiko
著者 鷲見 育亮
著者(別表記) Sumi Yasuaki
著者 副井 裕
著者(別表記) Fukui Yutaka
Abstract

In this paper, an automatic counting of chlorella using neural network(NN) is described. In the culture of marine chlorella, it is necessary to count the number in order to understand the condition of increase. The automatic counting method of chlorella, using image processing and NN, was proposed. This NN is three-layerd and uses Back Propogation as learning method. This paper describes the study on the increase of the differential learning data of the NN on the recognition rate of chlorella. Experiments are the first case on the increase or decrease of the number of the usual learning data, the second case on the increase or decrease of the number of the new learning data, and the third case on learning using the extracted data from the counted images in experiment.

Subject
000 General works
Keyword
chlorella counting
image processing
neural network
learning
Publish Date
2003-01-31
Publication Title
津山工業高等専門学校紀要
Publication Title Alternative
Bulletin of Tsuyama National College of Technology
Volume
44
Start Page
29
End Page
35
ISSN
0287-7066
NCID
AN00149351
Publisher
津山工業高等専門学校
Official URL
Contents Type
Departmental Bulletin Paper
language
Japanese
File Version
publisher