北京生物醫(yī)學(xué)工程

支持向量機(jī)在肺結(jié)節(jié)CT圖像中的應(yīng)用

Application of support vector machine for pulmonary nodules in CT image

作者：王晶晶孫濤趙楓朝李霞蔡博文朱曉萌郭秀花

單位：首都醫(yī)科大學(xué)公共衛(wèi)生學(xué)院(北京100069)

關(guān)鍵詞： CT圖像；肺結(jié)節(jié)；Curvelet轉(zhuǎn)換；紋理提取；支持向量機(jī)

分類號(hào)：

出版年·卷·期（頁碼）：2013·32·5（528-530）

摘要：

目的探討基于孤立性肺結(jié)節(jié)建立支持向量機(jī)預(yù)測(cè)模型效果，提高肺癌的早期診斷率。方法對(duì)收集的55例患者的三正交位的4135張肺結(jié)節(jié)CT圖像，應(yīng)用Curvelet變換進(jìn)行紋理提取，對(duì)提取的476個(gè)特征值應(yīng)用支持向量機(jī)進(jìn)行良惡性分類并預(yù)測(cè)，應(yīng)用符合率、敏感度和特異度對(duì)預(yù)測(cè)結(jié)果進(jìn)行評(píng)估。結(jié)果 CT圖像符合率為78.0%，敏感度為88.6%，特異度為24.0%。結(jié)論 Curvelet轉(zhuǎn)換提取三正交位肺結(jié)節(jié)紋理，用支持向量機(jī)建立預(yù)測(cè)模型，一定程度上有助于早期發(fā)現(xiàn)和診斷肺癌。

Objective To evaluate the prediction results of solitary pulmonary nodules using support vector machine model in order to improve the detection and diagnosis of early-stage lung cancer.Methods We collected 4135 CT images of benign or malignant solitary pulmonary nodules in three dimensions from 55 patients.Four hundred and seventy-six Curvelet transform textural features were used as parameters to establish support vector machine model，and the classification consistency，sensitivity and specificity were used to evaluate the forecast results.Results The classification consistency，sensitivity and specificity for the model were 78.0%，88.6% and 24.0%，respectively.Conclusions Based on Curvelet transform to extract textural features，support vector machine can improve the diagnosis of early-stage lung cancer to some extent.

參考文獻(xiàn)：

［1］Jemal A，Bray F，Center MM，et al.Global cancer statistics［J］.CA Cancer J Clin,2011,61(2):69-90.
［2］郭鋒杰,范亞光,喬友林,等.HPV和肺癌關(guān)系的研究進(jìn)展［J］.中國肺癌雜志,2012,15(3):191-194.
Guo Fengjie，F(xiàn)an Yaguang，Qiao Youlin，et al.Study advance of relationship between HPV and lung cancer［J］.Chinese Journal of Lung Cancer，2012,15(3):191-194.
［3］韋春暉.肺癌早期診斷進(jìn)展［J］.臨床肺科雜志,2010,15(8):1136-1138.
Wei Chunhui.The early diagnosis progress of lung cancer［J］.Journal of Clinical Pulmonary Medicine，2010,15(8): 1136-1138.
［4］戴世明.CT技術(shù)在早期肺癌診斷中的應(yīng)用［J］.臨床肺科雜志,2012,17(2):330-331.
Dai Shiming.The application of CT technology in diagnosis of early stage lung cancer［J］.Journal of Clinical Pulmonary Medicine，2012,17(2):330-331.
［5］付杰，董云.50例周圍型肺癌的CT診斷價(jià)值分析［J］.中國醫(yī)藥科學(xué)，2011,1（21）：98-99.
Fu Lei, Dong Yun. The analysis on CT technology in dignosis of 50 cases of peripheral lung cancer［J］.China Medicine and Pharmacy,2011,1(21):98-99.
［6］Wang H，Guo XH，Jia ZW，et al.Multilevel binomial logistic prediction model for malignant pulmonary nodules based on texture features of CT image［J］.European Journal of Radiology,2010,74(1):124-129.
［7］吳海豐,劉韞寧,郭秀花,等.Curvelet變換在醫(yī)學(xué)圖像處理中的應(yīng)用現(xiàn)況［J］.北京生物醫(yī)學(xué)工程,2010,29(4):432-435.

Wu Haifeng，Liu Yunning，Guo Xiuhua，et al.The application status of Curvelet transformation in medical image processing［J］.Beijing Biomedical Engineering，2010,29(4):432-435.
［8］Dettori L，Semler L.A comparison of wavelet，ridgelet，and curvelet-based texture classification algorithms in computed tomography.Computers in Biology and Medicine,2009，(37): 486-498.
［9］Chen M，Zhang JX.Color segmentation of nuclei of blood cell using support vector machine［J］.Journal of Optoelectronics Laser，2006，17(4): 479-483.
［10］Zheng Z，Zhang YX，Hu YX.Investigation of eye gaze based on independent component analysis and support vector machine［J］.Journal of Optoelectronics Laser,2007,18(7): 491-494.
［11］王瓛,郭秀花,李坤成,等.良惡性肺小結(jié)節(jié)CT圖像基于灰度共生矩陣10種紋理特征研究［J］.北京生物醫(yī)學(xué)工程,2008,27(6):561-564,608.
Wang Huan，Guo Xiuhua，Li Kuncheng，et al.CT Images’10 texture features of small solitary pulmonary nodules patients using gray level co-occurrence matrix［J］.Beijing Biomedical Engineering，2008，27(6):561-564,608.
［12］吳海豐,劉韞寧,孫濤,等.基于Curvelet變換提取肺結(jié)節(jié)圖像紋理特征構(gòu)建BP神經(jīng)網(wǎng)絡(luò)［J］.北京生物醫(yī)學(xué)工程,2011，30(5): 471-473.
Wu Haifeng，Liu Yunning，Sun Tao，et al.Classification of malignant and benign pulmonary nodules in CT image based on Curvelet transformation［J］.Beijing Biomedical Engineering，2011，30(5): 471-473.
［13］Wu HF，Sun T，Guo XH，et al.Combination of radiological and gray level co-occurrence matrix textural features used to distinguish solitary pulmonary nodules by computed tomography［J/OL］.Journal of Digital Imaging. DOI: 10.1007/s10278-012-9547-6.

服務(wù)與反饋：

【文章下載】【加入收藏】

提示：您還未登錄，請(qǐng)登錄！點(diǎn)此登錄

51黑料吃瓜在线观看,51黑料官网|51黑料捷克街头搭讪_51黑料入口最新视频