Liver Cancer Identification Based on Improved SVM Model

摘要：

This paper proposes a novel liver cancer identification method based on improved SVM method. First, multiply features are extracted from ROI. Then GAs are used to feature selection. The classification model is built based on features selected. Moreover, improved SVM Universal network model is incorporated in the liver cancer classifier in order to avoid the disturbance of Liver cyst and angeioma, which consists of three SVMs. The experimental results show that the proposed method overcomes the shortcomings of traditional SVM method for liver cancer identification.

作者: JIANG Hui-Yan TAO Chun-Yan ZHANG Xi-Yue MAO Ke-Ming

作者单位: Multimedia Medical Information Technology Laboratory, Software College, Northeastern University, Shenyang, Liaoning, China

会议类型: 国际会议

会议名称: Yangtze River 2009 International Conference on Medical Imaging Physics & The 5th National Annual Meeting of Medical Imaging Physics(长江2009国际医学影像物理和工程大会暨第五届中国医学影像物理学术年会ICMIP2009)

会议地点: 南京

会议语种:英文

页码: 156-160

在线出版日期: 2009-10-23（万方平台首次上网日期，不代表论文的发表时间）