基于特性模型与神经网络的乳腺图像肿块自动检测技术

徐伟栋; 刘伟; 厉力华; 夏顺仁; 马莉; 邵国良; 张娟

doi:10.3724/SP.J.1146.2008.00677

基于特性模型与神经网络的乳腺图像肿块自动检测技术

doi: 10.3724/SP.J.1146.2008.00677

基金项目:

国家杰出青年基金(60788101)，国家自然科学基金(60705016，60775016)和浙江省自然科学基金(Y1080740)和浙江省科技计划重大攻关项目(2006C14026)资助课题

计量
- 文章访问数: 3216
- HTML全文浏览量: 61
- PDF下载量: 1068
- 被引次数: 0
出版历程
- 收稿日期: 2008-05-30
- 修回日期: 2009-03-09
- 刊出日期: 2009-07-19

Automatic Detection of the Masses in the Mammograms Using Characteristic Modeling and Neural Networks

摘要

摘要: 钼靶X线摄影是最常用的乳腺癌早期诊断手段。该文针对乳腺图像中的肿块提出了一种基于特性模型与神经网络的计算机辅助诊断技术。它首先建立两种特性模型分别描述脂肪组织和腺体组织中的肿块；然后对脂肪中的肿块采用迭代阈值法进行检测，对腺体中的肿块采用小波域黑洞检索法进行标记；接着采用一种基于Canny算子和能量场约束以及ANFIS控制的填充膨胀方法分割疑似肿块；最后使用一种MLP分类器剔除假阳性。实验结果表明，该算法在面对特性迥异的多种肿块时可取得较高的检测精度，并保证较低的假阳性率。
- 乳腺X线图像;计算机辅助诊断;肿块;ANFIS;MLP
Abstract: Mammography is a conventional early detection method for breast cancer. A novel Computer-Aided Diagnosis (CAD) method for the masses is proposed in this paper. Two characteristic models are built up to represent the masses with various backgrounds, and iterative thresholding is carried out to detect the masses in the fatty tissue; however, black-hole detection of wavelet-domain is applied to label the masses in the dense tissue. Filling dilation based on ANFIS controller, Canny detector and the energy field constraint is used to segment the suspicious masses, and MLP-based classifier is applied to suppress the false positives. The experiments validate that the proposed algorithm gets high detection precision, as well as low false positive rate.

HTML全文

参考文献(1)

Polakowski W E, Cournoyer D A, and Rogers S K, et al..Computer-aided breast cancer detection and diagnosis ofmasses using difference of Gaussians and derivative-basedfeature saliency[J].IEEE Transactions on Medical Imaging.1997, 16(6):811-819[2]Zheng L and Chan A K. An artificial intelligent algorithm fortumor detection in screening mammogram[J].IEEETransactions on Medical Imaging.2001, 20(7):559-567[3]Sampat M P and Bovik A C. Detection of spiculated lesionsin mammograms. 25th IEEE Annual InternationalConference of the Engineering in Medicine and BiologySociety, Cancun, Mexico, September 2003: 810-813.[4]Abdel-Dayem A R and El-Sakka M R. Fuzzy entropy baseddetection of suspicious masses in digital mammogram images.27th Annual Conference of IEEE Engineering in Medicineand Biology, Shanghai, China, September 2005: 4017-4022.[5]Cascio D, Fauci F, and Magro R, et al.. Mammogramsegmentation by contour searching and mass lesionsclassification with neural network[J].IEEE Transactions onNuclear Science.2006, 53(5):2827-2833[6]徐伟栋, 王小英, 夏顺仁, 严勇. 基于模型的乳腺X 线图像胸肌分割算法研究. 浙江大学学报( 工学版), 2005, 39(3):427-432.Xu W D, Wang X Y, Xia S R, and Yan Y. Study onmodel-based pectoral-muscle segment algorithm inmammograms. Journal of Zhejiang University (EngineeringScience), 2005, 39(3): 427-432.[7]Xu W D, Xia S R, and Duan H L. Segmentation of mass inmammograms using a novel intelligent algorithm[J].International Journal on Pattern Recognition and ArtificialIntelligence.2006, 20(2):255-270[8]Li L H, Wu Z B, and Salem A F. Computerized analysis oft5 density effect on missed cancer detection in digitalmammography[J].Computerized Medical Imaging and Graphics.2006, 30(5):291-297[9]Xiao M, Xia S R, and Wang S W. Geometric active contourmodel with color and intensity priors for medical imagesegmentation. 27th IEEE Annual International Conference ofthe Engineering in Medicine and Biology Society, Shanghai,China, September 2005: 6496-6499.[10]张智星, 孙春在, 水谷英二. 张平安, 高春华, 高峰. 神经-模糊和软计算. 第1 版, 西安: 西安交通大学出版社, 2000:238-287.Jang J S R, Sun C T, Mizutani E, Zhang P A, Gao C H, andGao F. Neuro-Fuzzy and Soft Computing. 1st Edition, Xian:Xian Jiaotong University Press, 2000: 238-287.[11]Rangayyan R M, El-Faramawy N M, and Desautels J E L, etal.. Measures of acutance and shape for classification ofbreast tumors[J].IEEE Transactions on Medical Imaging.1997,16(6):799-810[12]Mudigonda N R, Rangayyan R M, and Desautels J E L.Detection of breast masses in mammograms by density slicingand texture flow-field analysis[J].IEEE Transactions onMedical Imaging.2001, 20(12):1215-1227

施引文献

资源附件(0)

访问统计