상세정보
학술대회
A Support Vector Machine Based Classifier to Extract Abnormal Features from Breast Magnetic Resonance Images
- 발행일
- 201210
- 출처
- Research in Applied Computation Symposium (RACS) 2012, pp.158-152
- 협약과제
- 12PR3800, 전자파 이용 조기진단 고정밀 MT 시스템 개발, 전순익
- 초록
- Magnetic resonance imaging (MRI) is one of the high quality technologies to detect the breast cancer. This study proposes a new framework to extract abnormal features in Magnetic Resonance (MR) images by concentrating on the key aspect of the features: generating a unique input sequence to apply the Support Vector Machine (SVM) classifier. The main contribution of the proposed approach is the improvement of an accuracy in identifying abnormal features using SVM classifier. This approach is also less sensitive to noise in detecting the breast cancer. In order to evaluate the improved performance of the proposed SVM classifier, the results of traditional Decision Tree (DT) classifier has been compared with that of SVM. Copyright 2012 ACM.
- KSP 제안 키워드
- Breast Cancer(BC), Decision Tree(DT), Magnetic resonance(MR), Magnetic resonance images, Support Vector Machine (SVM) classifier, improved performance