期刊名称：International Journal of Computer Science & Technology
印刷版ISSN：2229-4333
电子版ISSN：0976-8491
出版年度：2011
卷号：2
期号：3
出版社：Ayushmaan Technologies
摘要：The techniques of pattern recognition to classify the characteristics of images have been based on the similarities of certain features. To detect possible breast cancer, some bioinformatics inspired thermogram features are to be used to classify the associated stages. In this research we specifically applied Fuzzy Inference System method to a total of 60 thermograms of normal breasts as well as of those in early and advance cancer. In addition four standard image processing were also used to enhance the detection capabilitity. The results show that the Fuzzy Inference System method alone has the accuracies of the average RMSE for mean value, standard deviation, entropy, skewness, and kurtosis after 100 epochs of learning fuzzy are 0,1628; 0,24; 0,36; 2,35; 0,4481 respectively. With additional wiener filtering, histogram equalization, and region growing on the original images, the average RMSE have decreased 0,0298; 0,147; 0,3424; 0,0322; 0,3476 respectively.
关键词：Thermogram images; Fuzzy Inference System; wiener filtering;histogram equalization.security.y.
