| The in-depth contents |
| Publication year |
2006 |
| PDF |
Download |
| Title |
DEVELOPMEKT OF AUTOMATIC DIAGKOSTIC SYSTEM FOR DETERIORATED MEMBRANE MATERIALS with Image Analysis and Image Recognition |
| Author |
Jun YANAGIMURO,Wonyong KOO,Atsushi TAKIZAWA,Naiki KATOH,Hiroshi TOYODA,Jun FUJIWARA,Kenshi ODA
|
| Summary |
Digital photographs taken with the microscope are used practically to diagnose deteriorated membrane materials. In this paper we present a new automated diagnostic system utilizing image analysis and image recognition instead of checking with human eyes. We first extract from a digital image statistics of co]or information and features of the objects that image along with deterioration such as cracks on the polyethylene resin coating and exposed fabrics. We then use these features as explanatory variables for building prediction models which are constructed based on multiple linear regression and a certain type of date mining method which combines multiple linear regression and decision tree. We have carried out comparison of prediction models through cross validation test. As the result of the experimental test, our prediction model exhibited more than 90% accuracy in coefficient of correlation.
|