| 講演抄録/キーワード |
| 講演名 |
2009-01-20 14:30
[ポスター講演]Automated Identification of Clustered Microcalcifications in Digital Mammography ○Yi-Jhe Huang・Da-Chuan Cheng(China Medical Univ.)・Po-Pang Tsai・Wu-Chung Shen(China Medical Univ. Hospital) MI2008-153 |
| 抄録 |
(和) |
Computer-aided diagnosis (CAD) system for early diagnosis of breast cancer has been developed as a useful tool for radiologists since many years. However, the commercial product can identify part of lesion in detecting microcalcification. In this study, we used gray level co-occurrence matrix (GLCM) to extract texture features. Fourteen features are analyzed and some of them are selected which are more significant for classification. Afterwards, support vector machine (SVM) has been used to perform a nonlinear classification. The accuracy of proposed system is 90.44 %, the sensitivity is 73.53 %, and the specificity is 93.82 %. |
| (英) |
Computer-aided diagnosis (CAD) system for early diagnosis of breast cancer has been developed as a useful tool for radiologists since many years. However, the commercial product can identify part of lesion in detecting microcalcification. In this study, we used gray level co-occurrence matrix (GLCM) to extract texture features. Fourteen features are analyzed and some of them are selected which are more significant for classification. Afterwards, support vector machine (SVM) has been used to perform a nonlinear classification. The accuracy of proposed system is 90.44 %, the sensitivity is 73.53 %, and the specificity is 93.82 %. |
| キーワード |
(和) |
Breast cancer / Microcalcifications (MCs) / Computer-aided diagnosis (CAD) / gray level co-occurrence matrix (GLCM) / support vector machine (SVM) / / / |
| (英) |
Breast cancer / Microcalcifications (MCs) / Computer-aided diagnosis (CAD) / gray level co-occurrence matrix (GLCM) / support vector machine (SVM) / / / |
| 文献情報 |
信学技報, vol. 108, no. 385, MI2008-153, pp. 431-434, 2009年1月. |
| 資料番号 |
MI2008-153 |
| 発行日 |
2009-01-12 (MI) |
| ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
| PDFダウンロード |
MI2008-153 |