Detection Method of Micro-calcification Clusters Based on Classification and Regression Decision Tree
Given that the micro-calcification clusters in early breast cancer X-ray pictures are minimal and irregular with differentiated shapes and distributions as well as the unsatisfactorily low contrast ratio, micro-calcification clusters of small sizes and the unsatisfactorily low contrast ratio tend to be easily ignored or misdiagnosed by doctors. Based on the integration of Wavelet Transform and mathematical graphics, Classification and Regression Decision Tree (CART) has been used to training for the classification. With CART, pixels of tumor area and pixels non-tumor area can be thus distinguished so as to improve the detection rate of true positivity while reduce the diagnosis of false positivity. In addition to the simple operations, this method presents relatively higher detection efficiency and is proved to be efficient in significantly reducing the false positivity.
mammary gland X-ray pictures CART algorithm micro-calcification detection Wavelet Transform
Kong Ying Zhao Jie
School of Information and Electronic Engineering Zhejiang University of Science and Technology Hangz School of Experimentation and Practice Training Management Center Zhejiang Police Vocational Academy
国际会议
成都
英文
1225-1228
2010-12-17（万方平台首次上网日期，不代表论文的发表时间）