基于支持向量机的超声乳腺肿瘤图像计算机辅助诊断系统.pdfVIP

19 1 2008 1 Journal of Optoelectronics Laser Vol. 19 No. 1 Jan. 2008 基于支持向量机的超声乳腺肿瘤图像计算机辅助诊断 * 系统* * * 李晓峰 , 沈 毅 (, 150001) :16 26 , , 93.03%,94.30%, 91.59%, 92. 80%, 93.33%, AUC0.9669,[ 1] :; ; ; ; :TP391.4 :A :2008 Support vector machinesbasedcomputeraideddiagnosis systemof breast tumor withul trasound images * * I Xiaofeng , SHEN Yi ( ab of Detection technology and automatic equipment, Control Science and Engineering, Harbin Institute of Tech nology, Harbin 150001, China) Abstract:A computeraided diagnosis system of breast tumor with ultrasound images based on isproposed. Based on the cri terions used by physicians, total twentysix features about textures and geometric feature including autocorrelation coeffi cients, mean squaredeviation coefficients,and compact ratios are extracted.Thesefeatures reflect theessential differences be tween malignant breast tumors and benign tumors, and concern about the characters of ultrasound images as the coherent images. Experiment results show that the classification accuracy is 93.03%, sensitivity is 94. 30%, specificity is 91. 59%, the positive predictive value is 92. 80%, negative predictive value is 93. 33%, the areas under the ROC curve is0.9669. Keywords: support vector machines; medical ultrasound image; computeraided diagnosis system; wavelet trans form; feature extraction ,, 1 [2] , 3554 , , , , , ,, , , 16 [ 3