The Paper's Abstract of The 3rd
International Conference on Imaging Science and Hardcopy

Image Recognition System using Hicrarchicl FuzzyFeature Representation

May 26-29,1998
Chongqing China

C. T. Chen and C. J. Hwang

Department of Computer Engineering & Science Yuan Zhi University
135 Yuan-Tung road, Chung-Li 320, Taiwan

    Image recognition is one of the most interested field for academic research. The elementary work of image recognition is to abstract some features from the patterns and then classified into a number of categories. However, uncertainty in computer vision may arise at various levels. It could occur in the raw sensor input, such as uneven illumination, and extends all the way through intermediate and higher levels. Hence a recognition system must have sufficient flexibility for representation of uncertainty. 
      We employ the fuzzy set theory to solve the uncertainty problems in image recognition. Firstly, wr designed a new feature extraction method. Also, we used fuzzy linguistic variable to represent those features, and used a hierarchical feature marching method. At the end, we simulate a ship recognition to show the results.

中国复印科学与工程学会
Reprographic Scientists and Engineers Society of China 
7 Chang-Tu Rd.Tianjin300131,China  Tel:(22)26669189 Fax: (22)26669189 
E-mail: rsesc@public.tpt.tj.cn
.