基于多示例学习的浅表器官超声图像分类方法研究
|
摘要
癌症严重威胁着人类的健康，其成功治愈依赖于病变的早期发现与治疗。超声成像以其无创、经济、便捷、无辐射等特点，已经成为浅表器官病变早期检测的首选方式。但是对超声图像的正确判读依赖于医生的知识水平和临床经验，具有较强的主观性。计算机辅助诊断技术的引入，为临床医生提供了客观、量化的病变信息和辅助决策信息，有助于消除由于医生主观因素造成的误诊，是临床医学研究的热点问题，对提高疾病的诊断准确率具有重要意义。目前基于超声图像的计算机辅助诊断系统主要存在两个问题：一是由于图像质量导致病变的精确定位相当困难，定位误差会对特征的准确提取与分类性能造成影响；二是缺乏在定位不精确前提下的分类方法研究。
本文针对以上问题，对超声图像计算机辅助诊断系统中不依赖于病变精确定位的分类方法进行了研究，提出将超声图像分类问题转化为多示例学习问题。本文考虑样本分布情况，提出了局部加权的Citation-kNN算法；考虑到超声图像存在良恶性特征交叉现象，提出了示例空间向概念空间映射的多示例学习方法；提出了评价弹性超声图像的量化指标，针对弹性图像和B超图像的特点，提出全局特征与局部特征相结合的多示例学习方法。主要工作包括以下三方面内容：
一、提出局部加权的Citation-kNN算法。对超声图像进行划分，其整体作为示例包，子区域作为示例，将超声图像分类问题转化为多示例学习问题。在传统Citation-kNN算法投票集合构造的基础上，进一步考虑样本的空间分布，对投票者采用距离加权和离散度加权，并对不同加权方式进行组合，在乳腺超声图像库和标准测试数据集上进行了实验并取得了良好的效果。
二、提出局部特征与多示例学习相结合的超声图像分类方法。超声图像中非病变组织会对病变的准确分类造成一定的干扰，对病变的定位将有助于提高分类的准确性。在对病变初步定位的基础上，对其进行划分，采用局部纹理特征对病变进行描述，将该定位区域作为示例包，其子区域作为示例。考虑到超声图像存在良恶性特征混叠的现象，传统的多示例学习定义不再适用。采用聚类方法获取概念空间，将示例包向概念空间投影，在概念空间构造分类器。在乳腺超声图像库上的实验表明，相比于其他方法，本文所提出的方法具有更高的分类准确性。
三、提出评价甲状腺弹性超声图像的定量指标，提出全局特征与局部特征相结合的多示例学习方法。通过对弹性超声图像的分析，提出可用于评价弹性图像的量化指标，与目前临床上评价弹性图像的方法相比，具有更高的准确性和可靠性。通过对弹性图像特征的分析，提出弹性图像的整体特征相比局部特征更有利于良恶性病变的区分。在对病变初步定位的基础上，将弹性图像的全局特征与B超图像的局部特征相结合，采用多示例学习方法进行分类，取得了良好的效果。
本文在超声图像分类方面的研究工作，较好地解决了病变定位不精确情况下的特征提取与分类问题，对超声图像计算机辅助诊断系统的研究具有积极的意义。
Cancer is a serious threat to human health. Its successful cure is dependent onearly detection and treatment. Ultrasound imaging is noninvasive, economic,convenient, no radiation and so on. It has become the preferred method for earlydetection of superficial organ diseases. However, correct interpretation for theultrasound images is dependent on the physician's knowledge level and clinicalexperience, has strong subjectivity. Computer aided diagnosis technology canprovides objective, quantification and decision making information about lesion.It is helpful to eliminate misdiagnosis in clinical practice due to subjective factors.It has become a hot issue in the medical domain. Moreover, it has importantsignificance for improving diagnosis accuracy. At present, there are twoproblems in the computer aided diagnosis system for ultrasound images. One isvery difficult to position lesions precisely due to the image quality. The error canaffect feature extraction and classification performance. The other is the lack ofstudy for classification method under the inaccurate positioning condition.
To solve above problems, this dissertation study the classification method ofcomputer-aided diagnosis system for ultrasound images, which do not rely onlesions precisely positioning. It proposes to convert the classification problem ofultrasound image to multiple-instance learning problem. It proposes locallyweighted Citation-kNN algorithm based on sample distribution. There existscrossover of benign and malignant lesions on ultrasound images. Thephenomenon is taking into account and the multiple-instance learning method,which based on mapping sample space to concept space, is proposed. Thedissertation proposes a quantitative criterion for evaluating elastogram. Based onthe characteristics of the B-mode images and elastogram, it proposes amultiple-instance learning method which combining global and local features.The main work includes the following three aspects:
1. Locally weighted Citation-kNN algorithm is proposed. By dividing theultrasound image, the whole image is a bag and sub-regions are instances. Thenthe classification problem of ultrasound image can convert to a multiple-instancelearning problem. By further considering spatial distribution of samples, the voter weighted according to distance and dispersion based on traditionalCitation-kNN’s voting set. The different weighting methods are combined andtested. The good results are achieved when experimented on breast ultrasoundimage database and benchmark for multiple-instance learning.
2. The classification method for ultrasound images which combining localfeatures and multiple-instance learning is proposed. The non-diseased tissue caninterfere with accurate classification of lesions. To position lesion will help toimprove the classification accuracy. The lesion is roughly located as region ofinterest (ROI) and the ROI is divided. Local texture features are used to describethe ROI. The ROI is a bag and sub-regions are instances. The traditionaldefinition of multiple-instance learning is no longer applicable due to thefeatures’ overlapping between benign and malignant lesion. Clustering method isused to construct the concept space. The bag is projected to the concept space.The classifier is trained on the concept space. The experiments show that theproposed method has higher classification accuracy on the ultrasound imagedatabase.
3. The quantitative criterion to evaluate thyroid elastogram is proposed. Amultiple-instance learning method is proposed which combining global and localfeatures. A quantitative indicator for evaluating elastogram is presented byanalyzing elastogram features. It has higher accuracy and reliability comparedwith current clinical evaluation methods. For the elastogram, the global featuresare more discriminable for benign and malignant lesions than local features. Afterlocated the lesion roughly, the global features of elastogram and the local featuresof B-mode images are combined. The proposed multiple-instance learningmethod has achieved better classification accuracy than others.
The research work is mainly on classification of ultrasound image. It solves thefeatures extraction and classification problem when imprecisely positioning thelesion well. In addition, it has positive significance to the computer-aideddiagnosis system.
本文针对以上问题，对超声图像计算机辅助诊断系统中不依赖于病变精确定位的分类方法进行了研究，提出将超声图像分类问题转化为多示例学习问题。本文考虑样本分布情况，提出了局部加权的Citation-kNN算法；考虑到超声图像存在良恶性特征交叉现象，提出了示例空间向概念空间映射的多示例学习方法；提出了评价弹性超声图像的量化指标，针对弹性图像和B超图像的特点，提出全局特征与局部特征相结合的多示例学习方法。主要工作包括以下三方面内容：
一、提出局部加权的Citation-kNN算法。对超声图像进行划分，其整体作为示例包，子区域作为示例，将超声图像分类问题转化为多示例学习问题。在传统Citation-kNN算法投票集合构造的基础上，进一步考虑样本的空间分布，对投票者采用距离加权和离散度加权，并对不同加权方式进行组合，在乳腺超声图像库和标准测试数据集上进行了实验并取得了良好的效果。
二、提出局部特征与多示例学习相结合的超声图像分类方法。超声图像中非病变组织会对病变的准确分类造成一定的干扰，对病变的定位将有助于提高分类的准确性。在对病变初步定位的基础上，对其进行划分，采用局部纹理特征对病变进行描述，将该定位区域作为示例包，其子区域作为示例。考虑到超声图像存在良恶性特征混叠的现象，传统的多示例学习定义不再适用。采用聚类方法获取概念空间，将示例包向概念空间投影，在概念空间构造分类器。在乳腺超声图像库上的实验表明，相比于其他方法，本文所提出的方法具有更高的分类准确性。
三、提出评价甲状腺弹性超声图像的定量指标，提出全局特征与局部特征相结合的多示例学习方法。通过对弹性超声图像的分析，提出可用于评价弹性图像的量化指标，与目前临床上评价弹性图像的方法相比，具有更高的准确性和可靠性。通过对弹性图像特征的分析，提出弹性图像的整体特征相比局部特征更有利于良恶性病变的区分。在对病变初步定位的基础上，将弹性图像的全局特征与B超图像的局部特征相结合，采用多示例学习方法进行分类，取得了良好的效果。
本文在超声图像分类方面的研究工作，较好地解决了病变定位不精确情况下的特征提取与分类问题，对超声图像计算机辅助诊断系统的研究具有积极的意义。
Cancer is a serious threat to human health. Its successful cure is dependent onearly detection and treatment. Ultrasound imaging is noninvasive, economic,convenient, no radiation and so on. It has become the preferred method for earlydetection of superficial organ diseases. However, correct interpretation for theultrasound images is dependent on the physician's knowledge level and clinicalexperience, has strong subjectivity. Computer aided diagnosis technology canprovides objective, quantification and decision making information about lesion.It is helpful to eliminate misdiagnosis in clinical practice due to subjective factors.It has become a hot issue in the medical domain. Moreover, it has importantsignificance for improving diagnosis accuracy. At present, there are twoproblems in the computer aided diagnosis system for ultrasound images. One isvery difficult to position lesions precisely due to the image quality. The error canaffect feature extraction and classification performance. The other is the lack ofstudy for classification method under the inaccurate positioning condition.
To solve above problems, this dissertation study the classification method ofcomputer-aided diagnosis system for ultrasound images, which do not rely onlesions precisely positioning. It proposes to convert the classification problem ofultrasound image to multiple-instance learning problem. It proposes locallyweighted Citation-kNN algorithm based on sample distribution. There existscrossover of benign and malignant lesions on ultrasound images. Thephenomenon is taking into account and the multiple-instance learning method,which based on mapping sample space to concept space, is proposed. Thedissertation proposes a quantitative criterion for evaluating elastogram. Based onthe characteristics of the B-mode images and elastogram, it proposes amultiple-instance learning method which combining global and local features.The main work includes the following three aspects:
1. Locally weighted Citation-kNN algorithm is proposed. By dividing theultrasound image, the whole image is a bag and sub-regions are instances. Thenthe classification problem of ultrasound image can convert to a multiple-instancelearning problem. By further considering spatial distribution of samples, the voter weighted according to distance and dispersion based on traditionalCitation-kNN’s voting set. The different weighting methods are combined andtested. The good results are achieved when experimented on breast ultrasoundimage database and benchmark for multiple-instance learning.
2. The classification method for ultrasound images which combining localfeatures and multiple-instance learning is proposed. The non-diseased tissue caninterfere with accurate classification of lesions. To position lesion will help toimprove the classification accuracy. The lesion is roughly located as region ofinterest (ROI) and the ROI is divided. Local texture features are used to describethe ROI. The ROI is a bag and sub-regions are instances. The traditionaldefinition of multiple-instance learning is no longer applicable due to thefeatures’ overlapping between benign and malignant lesion. Clustering method isused to construct the concept space. The bag is projected to the concept space.The classifier is trained on the concept space. The experiments show that theproposed method has higher classification accuracy on the ultrasound imagedatabase.
3. The quantitative criterion to evaluate thyroid elastogram is proposed. Amultiple-instance learning method is proposed which combining global and localfeatures. A quantitative indicator for evaluating elastogram is presented byanalyzing elastogram features. It has higher accuracy and reliability comparedwith current clinical evaluation methods. For the elastogram, the global featuresare more discriminable for benign and malignant lesions than local features. Afterlocated the lesion roughly, the global features of elastogram and the local featuresof B-mode images are combined. The proposed multiple-instance learningmethod has achieved better classification accuracy than others.
The research work is mainly on classification of ultrasound image. It solves thefeatures extraction and classification problem when imprecisely positioning thelesion well. In addition, it has positive significance to the computer-aideddiagnosis system.
引文
[1] Siegel R, Naishadham MA, Jemal A. Cancer statistics,2012[J]. A CancerJournal for Clinicians.2012,62:10–29.
[2] DeSantis C, Siegel R, Bandi P, Jemal A. Breast cancer statistics,2011[J].A Cancer Journal for Clinicians.2011,61:409–418
[3] J. Ferlay, D.M. Parkin, E. Steliarova-Foucher. Estimates of cancerincidence and mortality in Europe in2008[J]. European Journal ofCancer.2010,46:765–781
[4]王建伟,李安华.乳腺影像报告和数据系统-超声版应用进展[J].中华医学超声杂志,2010,11(7):79-81
[5] B.Liu, H.D.Cheng, J.Huang, J.Tian, X.Tang, J.Liu, Fully automatic andsegementation-robust classification of breast tumors based on localtexture analysis of ultrasound images[J]. Pattern Recogniton,2010,43(1):280-298
[6] Frates MC, Benson CB, Charboneau JW, et al. Management of thyroidnodules detected at US: Society of Radiologists in Ultrasound consensusconference statement[J]. Radiology.2005.237(3):794-800
[7] Tan GH, Gharib H. Thyroid incidentalomas: management approaches tononpalpable nodules discovered incidentally on thyroid imaging[J].Annals of Internal Medicine.1997,126(3):226-231
[8] Sprague BL, Warren AS, Trentham-Dietz A. Thyroid cancer incidenceand socioeconomic indicators of health care access[J]. Cancer CausesControl.2008.19(6):585-593
[9] Erickson BJ, Bartholmai B. Computer-aided detection and diagnosis atthe start of the third millennium[J]. Journal of Digital Imaging.2002,15:59–68
[10] Chan HP, Doi K, Vyborny CJ, Schmidt RA, Metz CE, Lam KL, et al.Improvement in radiologists’ detection of clustered microcalcificationson mammograms: the potential of computer-aided diagnosis[J].Investigative Radiology1990,25:1102–1110
[11] Moberg K, Bjurstam N, Wilczek B, Rostgard L, Egge E, Muren C.Computer assisted detection of interval breast cancers[J]. EuropeanJournal of Radiology2001,39:104–110
[12] Li F, Arimura H, Suzuki K, Shiraishi J, Li Q, Abe H, et al. Computeraided diagnosis for detection of missed peripheral lung cancers on CT:ROC and LROC analysis[J]. Radiology2005,237:684–690
[13] Hirai T, Korogi Y, Arimura H, Katsuragawa S, Kitajima M, Yamura M,et al. Intracranial aneurysms at MR angiography: effect of computer aideddiagnosis on radiologists’ detection performance[J]. Radiology2005,237:605–610
[14] B.S. Garra et al. Improving the Distinction Between Benign andMalignant Breast Lesions: The Value of Sonographic TextureAnalysis[J]. Ultrasonic Imaging1993,15:267–285
[15] B. Sahiner, H. P. Chan, M. A. Roubidoux, L. M. Hadjiiski, M. A. Helvie,C. Paramagul, J. Bailey, A. V. Nees, and C. Blane. Malignant and benignbreast masses on3-D US volumetric images: Effect of computer-aideddiagnosis on radiologist accuracy[J]. Radiology2007,242:716–724
[16] Chen DR, Chang RF, Huang YL. Computer-aided diagnosis applied toUS of solid breast nodules by using neural networks[J]. Radiology1999,213:407–412
[17] K. Horsch, M. L. Giger, C. J. Vyborny, and L. A. Venta. Performance ofcomputer-aided diagnosis in the interpretation of lesions on breastsonography[J]. Acadamic Radiololgy2004,11(3):272–280
[18] Drukker K, Gruszauskas NP, Sennett CA, Giger ML. Breast UScomputer-aided diagnosis workstation: performance with a large clinicaldiagnostic population[J]. Radiology2008,248:392–397
[19] Shen WC, Chang RF, Moon WK. Computer aided classification systemfor breast ultrasound based on Breast Imaging Reporting and Data System(BI-RADS)[J]. Ultrasound in Medicine&Biology2007,33(11):1688–1698
[20] D. E. Maroulis, M. Savelonas, D. K. Iakovidis, S. A. Karkanis, and N.Dimitropoulos. Variable background active contour model forcomputeraided delineation of nodules in thyroid ultrasound images[J].IEEE Transactions on Information Technology in Biomedicine2007,11(5):537–543
[21] S. Tsantis, D. Cavouras, I. Kalatzis, N. Piliouras, N. Dimitropoulos, andG. Nikiforidis. Development Of A Support Vector Machine-Based ImageAnalysis System For Assessing The Thyroid Nodule Malignancy Risk[J].Ultrasound in Medicine&Biology2005,31:1451-1459
[22] Chang C, Chen S, Tsai M. Application of support-vector-machine-basedmethod for feature selection and classification of thyroid nodules inultrasound images[J]. Pattern Recognition2010,43:3494-3506
[23] Savelonas M, Maroulis D, Sangriotis M: A computer-aided system formalignancy risk assessment of nodules in thyroid US images based onboundary features[J]. Comput Methods and Programs in Biomedicine2009,96:25–32
[24] Stavros AT, Thickman D, Rapp CL. et al. Solid breast nodules: use ofsonography to distinguish between benign and malignantlesions[J]. Radiology1995,196:123-134
[25] Rahbar G, Sie AC, Hansen GC, et al. Benign versus malignant solidbreast masses: US differentiation[J]. Radiology1999,213:889–894
[26] B. Sahiner, Malignant and benign breast masses on3D US volumetricimages: effect of computer-aided diagnosis on radiologist accuracy[J].Radiology2007,242(3):716–724
[27] K. Horsch, M.L. Giger, L.A. Venta, C.J. Vyborny. Computerizeddiagnosis of breast lesions on ultrasound[J]. Medical Physics2002,29(2):157–164
[28] K. Mogatadakala, K. Donohue, C. Piccoli, F. Forsberg, Detection ofbreast lesion regions in ultrasound images using wavelets and orderstatistics[J]. Medical Physics2006,33(4):840–849
[29] P.Shankar, V.Dumane, T.George, C.Piccoli, J.Reid, F.Forsberg,B.Goldberg, Classification of breast masses in ultrasonic B scans usingNakagami and K distribution[J]. Physics in Medicine and Biology2003,48(14):2229–2240
[30] J. Segyeong, S.Y. Yoon, K.M. Woo, C.K. Hee, Computer-aided diagnosisof solid breast nodules: use of an artificial neural network based onmultiple sonographic features[J]. IEEE Transactions on Medical Imaging2004,23(10):1292–1300
[31] D.R. Chen, R.F. Chang, Y.L. Huang, Breast cancer diagnosis usingself-organizing map for sonography[J]. Ultrasound in Medicine andBiology2000,26(3):405–411
[32] K. Zheng, T.F. Wang, J.L. Lin, D.Y. Li, Recognition of breast ultrasoundimages using a hybrid method[C]. in: IEEE/ICME InternationalConference on Complex Medical Engineering,2007,640–643
[33] K. Drukker, M.L. Giger, K. Horsch, M.A. Kupinski, C.J. Vyborny, E.B.Mendelson, Computerized lesion detection on breast ultrasound[J].Medical Physics2002,29(7):1438–1446
[34] K. Drukker, M.L. Giger, C.J. Vyborny, E.B. Mendelson, Computerizeddetection and classification of cancer on breast ultrasound[J]. AcademicRadiology2004,11(5):526–535
[35] K. Drukker, D.C. Edwards, M.L. Giger, R.M. Nishikawa, C.E. Metz,Computerized detection and3-way classification of breast lesions onultrasound images[C]. in: Proceedings of SPIE, Medical Imaging2004,5370:.1034–1041
[36] K. Drukker, M. Giger, C. Metz, Robustness of computerized lesiondetection and classification scheme across different breast USplatforms[J]. Radiology2006,238(1):834–840
[37] D.R. Chen, W.J. Kuo, R.F. Chang, W.K. Moon, C.C. Lee, Use of thebootstrap technique with small training sets for computer-aided diagnosisin breast ultrasound[J]. Ultrasound in Medicine and Biology2002,28(7):897–902
[38] W. Kuo, R. Chang, D. Chen, C. Lee, Data mining with decision trees fordiagnosis of breast tumor in medical ultrasonic images[J]. Breast CancerResearch and Treatment2001,66(1):51–57
[39] R.F. Chang, W.J. Wu, W.K. Moon, D.R. Chen, Improvement in breasttumor discrimination by support vector machines and speckle-emphasistexture analysis[J]. Ultrasound in Medicine and Biology2003,29(5):679–686
[40] Y. Huang, K. Wang, D. Chen, Diagnosis of breast tumors with ultrasonictexture analysis using support vector machines[J]. Neural Computing&Applications2006,15(2):164–169
[41] Y.L. Huang, D.R. Chen, Support vector machines in sonographyapplication to decision making in the diagnosis of breast cancer[J].Clinical Imaging2005,29(3):179–184
[42] Y.L. Huang, D.R. Chen, Y.K. Liu, Breast cancer diagnosis using imageretrieval for different ultrasonic systems[C]. in: International Conferenceon Image Processing,2004,5:2598–2960
[43] W. Kuo, R. Chang, C. Lee, W. Moon, D. Chen, Retrieval technique forthe diagnosis of solid breast tumors on sonogram[J]. Ultrasound inMedicine and Biology2002,28(7):903–909
[44]汪源源,沈嘉琳,王涌,王怡.基于形态特征判别超生图像中乳腺肿瘤的良恶性[J].光学精密工程2006,14(2):333-340
[45]杨盈,汪天富,彭玉兰,李德玉,林江莉,罗燕.基于形态和灰度特征的乳腺肿瘤B超图像识别[J].中国医学影像技术2005,21(11):1758-1760
[46]张科宏，彭玉兰，李德玉，林江莉，罗燕，汪天富，蒋银宝.基于边界特征的乳腺肿瘤超声图像识别[J].生物医学工程杂志2006,23(6):1237-1240
[47]李晓峰，沈毅.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[J].光电子激光2008,19(1):115-119
[48] American Cancer Society. Breast Cancer Facts&Figures2005-2006.Retrieved Aug.26,2012at:http://www.cancer.org/docroot/STT/content/STT_1x_Breast_Cancer_Facts__Figures_2005-2006.asp
[49] Ophir J, Cespedes I, Ponnekanti H, Yazdi Y, Li X. Elastography: aquantitative method for imaging the elasticity of biological tissues[J].Ultrason Imaging.1991.13(2):111-34
[50] Rago T, Santini F, Scutari M, Pinchera A, Vitti P. Elastography: newdevelopments in ultrasound for predicting malignancy in thyroidnodules[J]. The Journal of Clinical Endocrinology&Metabolism.2007.92(8):2917-2922
[51] Fukunari N. More accurate and sensitive diagnosis for thyroid tumorswith elastography[R]. Medix Suppl.2007:16-19
[52] C.-S. H. Woo Kyung Moon, Wei-Chih Shen, Etsuo Takada, Ruey-FengChang, Juliwati Joe, Michiko Nakajima, Masayuki Kobayashi, Analysisof Elastographic and B-Mode Features at Sonoelastography for BreastTumor Classification[J]. Ultrasound in medicine and biology2009,35:1794-1802
[53] M. K. Shirley Selvan, S. Shenbagadevi and S. Suresh, Feature Extractionfor Characterization of Breast Lesions in Ultrasound Echography andElastography[J]. Journal of Computer Science2010,6:67-74
[54] F. Gorunescu. Data mining techniques in computer-aided diagnosis:Non-invasive cancer detection[J]. International Journal ofBioengineering, Biotechnology and Nanotechnology.2008,1(2):105-108
[55] A. Saftoiu, et al. Dynamic analysis of EUS used for the differentiation ofbenign and malignant lymph nodes[J]. Gastrointestinal endoscopy2007,66:291-300
[56] A. Saftoiu, et al. Neural network analysis of dynamic sequences of EUSelastography used for the differential diagnosis of chronic pancreatitisand pancreatic cancer[J]. Gastrointestinal endoscopy2008,68:1086-1094
[57] Moon WK, et al. Solid breast masses: classification with computer-aidedanalysis of continuous US images obtained with probe compression[J].Radiology2005,236:458-464
[58] Chen CJ, et al.2-D ultrasound strain images for breast cancer diagnosisusing nonrigid subregion registration[J]. Ultrasound in Medicine andBiology,2006,32:837-846
[59]哈章，李传富，周康源等，联合弹性特征的多特征乳腺肿瘤超声图像识别[J].北京生物医学工程2008,27(6):565-568
[60]哈章，李传富，王金萍等，基于灰阶超声序列图像的乳腺肿瘤良恶性判别[J].中国医疗器械杂志2008,32(3):186-189
[61] Dietterich T G, Lathrop R H, Lozano-Pérez T. Solving themultiple-instance problem with axis-parallel rectangles[J]. ArtificialIntelligence1997,89(1-2):31-71
[62] O. Maron and T. Lozano-Perez. A framework for multiple-instancelearning[C]. In Proc. NIPS,1998,10:570–576
[63] Q. Zhang and S.A. Goldman. EM-DD: An Improved Multiple-InstanceLearning Technique[C]. Advances in Neural Information ProcessingSystems2002,14:1073-1080
[64] J. Wang and J.-D. Zucker. Solving the multiple-instance problem: a lazylearning approach[C]. In Proceedings of the17th InternationalConference on Machine Learning, San Francisco, CA,2000,1119–1125
[65] S. Andrews, I. Tsochantaridis, T. Hofmann. Support vector machines formultiple-instance learning[C]. Advances in Neural InformationProcessing Systems, MIT Press, Cambridge, MA,2002,561–568
[66] P. Gehler, O. Chapelle, Deterministic annealing for multiple-instancelearning[C]. Proceedings of the Eleventh International Conference onArtificial Intelligence and Statistics,2007,123–130
[67] O.L. Mangasarian, E.W. Wild, Multiple instance classification viasuccessive linear programming[J]. Journal of Optimization Theory andApplications2008,137(3):555–568
[68] J.-D. Zucker and Y. Chevaleyre. Solving multiple-instance andmultiplepart learning problems with decision trees and decision rules.Application to the mutagenesis problem[C]. in Proc.14th Biennial Conf.Can. Soc.Comput. Studies of Intelligence, Ottawa, ON, Canada,2001,204–214.
[69] Zhou, Z.-H., Zhang, M.-L.: Neural networks for multi-instancelearning[C]. In: Proceedings of the International Conference onIntelligent Information Technology, Beijing, China2002,455–459
[70] Z.-H. Zhou and M.-L. Zhang. Solving multi-instance problems withclassifier ensemble based on constructive clustering[J]. Knowledge andInformation Systems2007,11(2):155–170
[71] Zhang M-L, Zhou Z-H. Adapting RBF neural networks to multi-instancelearning[J]. Neural Process Letter2006,23(1):1–26
[72] Y. Chen and J. Z. Wang. Image Categorization by Learning andReasoning with Regions[J]. Journal of Machine Learning Research2004,5:913-939
[73] Y. Chen, J. Bi, and J. Z. Wang. MILES: Multiple-instance learning viaembedded instance selection[J]. IEEE Transactions on Pattern Analysisand Machine Intelligence2006,28(12):1931–1947
[74] Maron O, Ratan A L. Multiple-instance learning for natural sceneclassification[C]. In: Proceedings of the15th International Conference onMachine Learning, Madison, WI,1998,341-349
[75] C. Yang and T. Lozano-Perez. Image Database Retrieval withMultiple-Instance Learning Techniques[C]. Proc. InternationalConference on Data Engineering,2000,233-243
[76] Zhou Z-H, Jiang K, Li M. Multi-instance learning based web mining[J].Applied Intelligence2005,22(2):135–147
[77] Liang Zhu, Bo Zhao, Yang Gao. Multi-class Multi-instance Learning forLung Cancer Image Classification Based on Bag Feature Selection[J].Fifth International Conference on Fuzzy Systems and KnowledgeDiscovery, FSKD2008,2:487–492
[78] Weidmann N, Frank E, Pfahringer B. A two-level learning method forgeneralized multi-instance problems[C]. In: Proceedings of the14thEuropean Conference on Machine Learning, Cavtat-Dubrovnik, Croatia,2003,468-479
[79] Fung, G., et al. Multiple instance learning for computer aided diagnosis.In Advances in neural information processing systems2007,19:425-432
[80] B. Krishnapuram, et al. Multiple-instance learning improves CADdetection of masses in digital mammography[C]. In Proceedings of the9th international workshop on Digital Mammography,2008,350–357
[81] D. Wu, J. Bi, K. Boyer, A min-max framework of cascaded classifier withmultiple instance learning for computer aided diagnosis[C]. in: ComputerVision and Pattern Recognition (CVPR),2009,1359–1366
[82] H.D. Cheng, J. Shan, W. Ju, Y. Guo, L. Zhang. Automated breast cancerdetection and classification using ultrasound images: a survey[J]. PatternRecognition2010,43(1):299–317
[83] Kotsiantis S: Supervised Machine Learning: A Review of ClassificationTechniques[J]. Informatica2007,31:249-268
[84] Dietterich, T.G., Jain, A., Lathrop, R. H.,&Lozano-Pérez, T. Acomparison of dynamic reposing and tangent distance for drug activityprediction[C]. Advances in Neural Information Processing Systems1994,6:216–223. San Mateo: Morgan Kaufmann
[85] Edgar, G.A. Measure, topology, and fractal geometry (3rd print)[M].1995Springer-Verlag.
[86] Cover, T., Hart, P. Nearest neighbor pattern classification[J]. IEEETransactions on Information Theory1967,13(1):21–27
[87] Atkeson C, Moore A, Schaal S. Locally weighted learning[J]. ArtificialIntelligence Review,1997,11(1):11–73
[88] Vapnik V., Bottou L. Local algorithms for pattern recognition anddependencies estimation[J]. Neural Computation.1993,5(6):893-909
[89] J.A. Noble, D. Boukerroui. Ultrasound image segmentation: a survey[J].IEEE Transactions on Medical Imaging2006,25,(8):987–1010
[90]宇传华，徐勇勇．ROC分析的基本原理[J]．中华流行病学杂志，1998,19(2-A)：413-415
[91] Fawcett, T. An introduction to ROC analysis[J]. Pattern RecognitionLetter2006,27:861874
[92] Wagner R F, Smith S W, Sandrik J M, Lopez H. Statistics of speckle inultrasound B-scans[J]. IEEE Transactions on Sonics and Ultrasonics.1983,30:156–163
[93] H.D. Cheng, X.H. Jiang, Y. Sun, J.L. Wang, Color image segmentation:advances and prospects[J]. Pattern Recognition2001,34(12):2259–2281
[94] Y.H. Guo, H.D. Cheng, J.H. Huang, J.W. Tian, W. Zhao, L.T. Sun, Y.X.Su, Breast ultrasound image enhancement using fuzzy logic[J].Ultrasound in Medicine and Biology2006,32(2):237–247
[95] Pal, S. and Dutta-Majumder, D. Fuzzy Mathematical Approach to PatternRecognition[M]. Halsted Press,1986
[96] Haralick RM, Shanmugam HK, Dinstein I. Textural features for imageclassification[J]. IEEE Trans Syst Man Cybern1973,3:610–621
[97] F.M.J. Frank, J.M. Thijssen, Characterization of echographic imagetexture by cooccurrence matrix parameters[J]. Ultrasound Med. Biol.1997,23:559–571
[98] Foulds J, Frank E: A review of multi-instance learning assumptions[J].Knowl Eng Rev2010,25(1):1–25
[99] Harris, J.R., Lippman, M.E., Morrow, M., and Osborne, C.K., Eds..Diseases of the Breast2nd Ed[M]. Philadelphia: Lippincott Willams&Wilkins2000
[100] Kohonen T. Self-organizing map[C]. Proc IEEE1990,78(9):1464–1480
[101] J. Vesanto and E. Alhoniemi. Clustering of the self-organizing map[J].IEEE Trans. Neural Networks2000,11:586–600
[102] D. Blei, A. Ng, and M. Jordan. Latent dirichlet allocation[J]. Journal ofMachine Learning Research,3:993–1022,2003
[103] Fei-Fei, L and P. Perona A Bayesian Hierarchical Model for LearningNatural Scene Categories[C]. Proc. of IEEE Computer Vision and PatternRecognition.2005,524–531
[104] K. Wagstaff, S. Rogers, and S. Schroedl, Constrained K-means clusteringwith background knowledge[C]. in Proc.8th Int. Conf. MachineLearning,2001,577–584
[105] V.Vapnik. The nature of statistical learning theory[M]. Springer, NewYork,1995
[106] Kotsiantis S: Supervised Machine Learning: A Review of ClassificationTechniques[J]. Informatica2007,31:249-268
[107]罗建文，白净.超声弹性成像的研究进展.中国医疗器械信息[J].2005,11(5)：23-31
[108]任新平，詹维伟，超声弹性成像在乳腺疾病中的应用，诊断学理论与实践[J].2007,6(2):166-168
[109] T. Shiina, et al. Clinical assessment of real-time, freehand elasticityimaging system based on the combined autocorrelation method[C]. inProc. IEEE Ultrason. Symp.,2003,664–667
[110] Garra BS, Cespedes EI, Ophir J, et al. Elastography of breast lesions:initial clinical results[J]. Radiology1997,202:79–86
[111] Itoh A, Ueno E, Tohno E, et al. Breast disease: clinical application of USelastography for diagnosis[J]. Radiology2006,239:341–350
[112] Curiel L, Souchon R, Rouvière O, Gelet A, Chapelon JY. Elastographyfor the follow-up of high-intensity focused ultrasound prostate cancertreatment: initial comparison with MRI[J]. Ultrasound in Medicine&Biology2005,31:1461–1468
[113] Miyanaga N, Akaza H, Yamakawa M, et al. Tissue elasticity imaging fordiagnosis of prostate cancer: a preliminary report[J]. International Journalof Urology2006,13:1514–1518
[114] Lyshchik A, Higashi T, Asato R, et al. Thyroid gland tumor diagnosis atUS elastography[J]. Radiology2005,237:202–211
[115] Kagoya R, Monobe H, Tojima H, et al. Utility of elastography fordifferential diagnosis of benign and malignant thyroid nodules[J].Otolaryngology-Head Neck Surgery2010,143:230–234
[2] DeSantis C, Siegel R, Bandi P, Jemal A. Breast cancer statistics,2011[J].A Cancer Journal for Clinicians.2011,61:409–418
[3] J. Ferlay, D.M. Parkin, E. Steliarova-Foucher. Estimates of cancerincidence and mortality in Europe in2008[J]. European Journal ofCancer.2010,46:765–781
[4]王建伟,李安华.乳腺影像报告和数据系统-超声版应用进展[J].中华医学超声杂志,2010,11(7):79-81
[5] B.Liu, H.D.Cheng, J.Huang, J.Tian, X.Tang, J.Liu, Fully automatic andsegementation-robust classification of breast tumors based on localtexture analysis of ultrasound images[J]. Pattern Recogniton,2010,43(1):280-298
[6] Frates MC, Benson CB, Charboneau JW, et al. Management of thyroidnodules detected at US: Society of Radiologists in Ultrasound consensusconference statement[J]. Radiology.2005.237(3):794-800
[7] Tan GH, Gharib H. Thyroid incidentalomas: management approaches tononpalpable nodules discovered incidentally on thyroid imaging[J].Annals of Internal Medicine.1997,126(3):226-231
[8] Sprague BL, Warren AS, Trentham-Dietz A. Thyroid cancer incidenceand socioeconomic indicators of health care access[J]. Cancer CausesControl.2008.19(6):585-593
[9] Erickson BJ, Bartholmai B. Computer-aided detection and diagnosis atthe start of the third millennium[J]. Journal of Digital Imaging.2002,15:59–68
[10] Chan HP, Doi K, Vyborny CJ, Schmidt RA, Metz CE, Lam KL, et al.Improvement in radiologists’ detection of clustered microcalcificationson mammograms: the potential of computer-aided diagnosis[J].Investigative Radiology1990,25:1102–1110
[11] Moberg K, Bjurstam N, Wilczek B, Rostgard L, Egge E, Muren C.Computer assisted detection of interval breast cancers[J]. EuropeanJournal of Radiology2001,39:104–110
[12] Li F, Arimura H, Suzuki K, Shiraishi J, Li Q, Abe H, et al. Computeraided diagnosis for detection of missed peripheral lung cancers on CT:ROC and LROC analysis[J]. Radiology2005,237:684–690
[13] Hirai T, Korogi Y, Arimura H, Katsuragawa S, Kitajima M, Yamura M,et al. Intracranial aneurysms at MR angiography: effect of computer aideddiagnosis on radiologists’ detection performance[J]. Radiology2005,237:605–610
[14] B.S. Garra et al. Improving the Distinction Between Benign andMalignant Breast Lesions: The Value of Sonographic TextureAnalysis[J]. Ultrasonic Imaging1993,15:267–285
[15] B. Sahiner, H. P. Chan, M. A. Roubidoux, L. M. Hadjiiski, M. A. Helvie,C. Paramagul, J. Bailey, A. V. Nees, and C. Blane. Malignant and benignbreast masses on3-D US volumetric images: Effect of computer-aideddiagnosis on radiologist accuracy[J]. Radiology2007,242:716–724
[16] Chen DR, Chang RF, Huang YL. Computer-aided diagnosis applied toUS of solid breast nodules by using neural networks[J]. Radiology1999,213:407–412
[17] K. Horsch, M. L. Giger, C. J. Vyborny, and L. A. Venta. Performance ofcomputer-aided diagnosis in the interpretation of lesions on breastsonography[J]. Acadamic Radiololgy2004,11(3):272–280
[18] Drukker K, Gruszauskas NP, Sennett CA, Giger ML. Breast UScomputer-aided diagnosis workstation: performance with a large clinicaldiagnostic population[J]. Radiology2008,248:392–397
[19] Shen WC, Chang RF, Moon WK. Computer aided classification systemfor breast ultrasound based on Breast Imaging Reporting and Data System(BI-RADS)[J]. Ultrasound in Medicine&Biology2007,33(11):1688–1698
[20] D. E. Maroulis, M. Savelonas, D. K. Iakovidis, S. A. Karkanis, and N.Dimitropoulos. Variable background active contour model forcomputeraided delineation of nodules in thyroid ultrasound images[J].IEEE Transactions on Information Technology in Biomedicine2007,11(5):537–543
[21] S. Tsantis, D. Cavouras, I. Kalatzis, N. Piliouras, N. Dimitropoulos, andG. Nikiforidis. Development Of A Support Vector Machine-Based ImageAnalysis System For Assessing The Thyroid Nodule Malignancy Risk[J].Ultrasound in Medicine&Biology2005,31:1451-1459
[22] Chang C, Chen S, Tsai M. Application of support-vector-machine-basedmethod for feature selection and classification of thyroid nodules inultrasound images[J]. Pattern Recognition2010,43:3494-3506
[23] Savelonas M, Maroulis D, Sangriotis M: A computer-aided system formalignancy risk assessment of nodules in thyroid US images based onboundary features[J]. Comput Methods and Programs in Biomedicine2009,96:25–32
[24] Stavros AT, Thickman D, Rapp CL. et al. Solid breast nodules: use ofsonography to distinguish between benign and malignantlesions[J]. Radiology1995,196:123-134
[25] Rahbar G, Sie AC, Hansen GC, et al. Benign versus malignant solidbreast masses: US differentiation[J]. Radiology1999,213:889–894
[26] B. Sahiner, Malignant and benign breast masses on3D US volumetricimages: effect of computer-aided diagnosis on radiologist accuracy[J].Radiology2007,242(3):716–724
[27] K. Horsch, M.L. Giger, L.A. Venta, C.J. Vyborny. Computerizeddiagnosis of breast lesions on ultrasound[J]. Medical Physics2002,29(2):157–164
[28] K. Mogatadakala, K. Donohue, C. Piccoli, F. Forsberg, Detection ofbreast lesion regions in ultrasound images using wavelets and orderstatistics[J]. Medical Physics2006,33(4):840–849
[29] P.Shankar, V.Dumane, T.George, C.Piccoli, J.Reid, F.Forsberg,B.Goldberg, Classification of breast masses in ultrasonic B scans usingNakagami and K distribution[J]. Physics in Medicine and Biology2003,48(14):2229–2240
[30] J. Segyeong, S.Y. Yoon, K.M. Woo, C.K. Hee, Computer-aided diagnosisof solid breast nodules: use of an artificial neural network based onmultiple sonographic features[J]. IEEE Transactions on Medical Imaging2004,23(10):1292–1300
[31] D.R. Chen, R.F. Chang, Y.L. Huang, Breast cancer diagnosis usingself-organizing map for sonography[J]. Ultrasound in Medicine andBiology2000,26(3):405–411
[32] K. Zheng, T.F. Wang, J.L. Lin, D.Y. Li, Recognition of breast ultrasoundimages using a hybrid method[C]. in: IEEE/ICME InternationalConference on Complex Medical Engineering,2007,640–643
[33] K. Drukker, M.L. Giger, K. Horsch, M.A. Kupinski, C.J. Vyborny, E.B.Mendelson, Computerized lesion detection on breast ultrasound[J].Medical Physics2002,29(7):1438–1446
[34] K. Drukker, M.L. Giger, C.J. Vyborny, E.B. Mendelson, Computerizeddetection and classification of cancer on breast ultrasound[J]. AcademicRadiology2004,11(5):526–535
[35] K. Drukker, D.C. Edwards, M.L. Giger, R.M. Nishikawa, C.E. Metz,Computerized detection and3-way classification of breast lesions onultrasound images[C]. in: Proceedings of SPIE, Medical Imaging2004,5370:.1034–1041
[36] K. Drukker, M. Giger, C. Metz, Robustness of computerized lesiondetection and classification scheme across different breast USplatforms[J]. Radiology2006,238(1):834–840
[37] D.R. Chen, W.J. Kuo, R.F. Chang, W.K. Moon, C.C. Lee, Use of thebootstrap technique with small training sets for computer-aided diagnosisin breast ultrasound[J]. Ultrasound in Medicine and Biology2002,28(7):897–902
[38] W. Kuo, R. Chang, D. Chen, C. Lee, Data mining with decision trees fordiagnosis of breast tumor in medical ultrasonic images[J]. Breast CancerResearch and Treatment2001,66(1):51–57
[39] R.F. Chang, W.J. Wu, W.K. Moon, D.R. Chen, Improvement in breasttumor discrimination by support vector machines and speckle-emphasistexture analysis[J]. Ultrasound in Medicine and Biology2003,29(5):679–686
[40] Y. Huang, K. Wang, D. Chen, Diagnosis of breast tumors with ultrasonictexture analysis using support vector machines[J]. Neural Computing&Applications2006,15(2):164–169
[41] Y.L. Huang, D.R. Chen, Support vector machines in sonographyapplication to decision making in the diagnosis of breast cancer[J].Clinical Imaging2005,29(3):179–184
[42] Y.L. Huang, D.R. Chen, Y.K. Liu, Breast cancer diagnosis using imageretrieval for different ultrasonic systems[C]. in: International Conferenceon Image Processing,2004,5:2598–2960
[43] W. Kuo, R. Chang, C. Lee, W. Moon, D. Chen, Retrieval technique forthe diagnosis of solid breast tumors on sonogram[J]. Ultrasound inMedicine and Biology2002,28(7):903–909
[44]汪源源,沈嘉琳,王涌,王怡.基于形态特征判别超生图像中乳腺肿瘤的良恶性[J].光学精密工程2006,14(2):333-340
[45]杨盈,汪天富,彭玉兰,李德玉,林江莉,罗燕.基于形态和灰度特征的乳腺肿瘤B超图像识别[J].中国医学影像技术2005,21(11):1758-1760
[46]张科宏，彭玉兰，李德玉，林江莉，罗燕，汪天富，蒋银宝.基于边界特征的乳腺肿瘤超声图像识别[J].生物医学工程杂志2006,23(6):1237-1240
[47]李晓峰，沈毅.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[J].光电子激光2008,19(1):115-119
[48] American Cancer Society. Breast Cancer Facts&Figures2005-2006.Retrieved Aug.26,2012at:http://www.cancer.org/docroot/STT/content/STT_1x_Breast_Cancer_Facts__Figures_2005-2006.asp
[49] Ophir J, Cespedes I, Ponnekanti H, Yazdi Y, Li X. Elastography: aquantitative method for imaging the elasticity of biological tissues[J].Ultrason Imaging.1991.13(2):111-34
[50] Rago T, Santini F, Scutari M, Pinchera A, Vitti P. Elastography: newdevelopments in ultrasound for predicting malignancy in thyroidnodules[J]. The Journal of Clinical Endocrinology&Metabolism.2007.92(8):2917-2922
[51] Fukunari N. More accurate and sensitive diagnosis for thyroid tumorswith elastography[R]. Medix Suppl.2007:16-19
[52] C.-S. H. Woo Kyung Moon, Wei-Chih Shen, Etsuo Takada, Ruey-FengChang, Juliwati Joe, Michiko Nakajima, Masayuki Kobayashi, Analysisof Elastographic and B-Mode Features at Sonoelastography for BreastTumor Classification[J]. Ultrasound in medicine and biology2009,35:1794-1802
[53] M. K. Shirley Selvan, S. Shenbagadevi and S. Suresh, Feature Extractionfor Characterization of Breast Lesions in Ultrasound Echography andElastography[J]. Journal of Computer Science2010,6:67-74
[54] F. Gorunescu. Data mining techniques in computer-aided diagnosis:Non-invasive cancer detection[J]. International Journal ofBioengineering, Biotechnology and Nanotechnology.2008,1(2):105-108
[55] A. Saftoiu, et al. Dynamic analysis of EUS used for the differentiation ofbenign and malignant lymph nodes[J]. Gastrointestinal endoscopy2007,66:291-300
[56] A. Saftoiu, et al. Neural network analysis of dynamic sequences of EUSelastography used for the differential diagnosis of chronic pancreatitisand pancreatic cancer[J]. Gastrointestinal endoscopy2008,68:1086-1094
[57] Moon WK, et al. Solid breast masses: classification with computer-aidedanalysis of continuous US images obtained with probe compression[J].Radiology2005,236:458-464
[58] Chen CJ, et al.2-D ultrasound strain images for breast cancer diagnosisusing nonrigid subregion registration[J]. Ultrasound in Medicine andBiology,2006,32:837-846
[59]哈章，李传富，周康源等，联合弹性特征的多特征乳腺肿瘤超声图像识别[J].北京生物医学工程2008,27(6):565-568
[60]哈章，李传富，王金萍等，基于灰阶超声序列图像的乳腺肿瘤良恶性判别[J].中国医疗器械杂志2008,32(3):186-189
[61] Dietterich T G, Lathrop R H, Lozano-Pérez T. Solving themultiple-instance problem with axis-parallel rectangles[J]. ArtificialIntelligence1997,89(1-2):31-71
[62] O. Maron and T. Lozano-Perez. A framework for multiple-instancelearning[C]. In Proc. NIPS,1998,10:570–576
[63] Q. Zhang and S.A. Goldman. EM-DD: An Improved Multiple-InstanceLearning Technique[C]. Advances in Neural Information ProcessingSystems2002,14:1073-1080
[64] J. Wang and J.-D. Zucker. Solving the multiple-instance problem: a lazylearning approach[C]. In Proceedings of the17th InternationalConference on Machine Learning, San Francisco, CA,2000,1119–1125
[65] S. Andrews, I. Tsochantaridis, T. Hofmann. Support vector machines formultiple-instance learning[C]. Advances in Neural InformationProcessing Systems, MIT Press, Cambridge, MA,2002,561–568
[66] P. Gehler, O. Chapelle, Deterministic annealing for multiple-instancelearning[C]. Proceedings of the Eleventh International Conference onArtificial Intelligence and Statistics,2007,123–130
[67] O.L. Mangasarian, E.W. Wild, Multiple instance classification viasuccessive linear programming[J]. Journal of Optimization Theory andApplications2008,137(3):555–568
[68] J.-D. Zucker and Y. Chevaleyre. Solving multiple-instance andmultiplepart learning problems with decision trees and decision rules.Application to the mutagenesis problem[C]. in Proc.14th Biennial Conf.Can. Soc.Comput. Studies of Intelligence, Ottawa, ON, Canada,2001,204–214.
[69] Zhou, Z.-H., Zhang, M.-L.: Neural networks for multi-instancelearning[C]. In: Proceedings of the International Conference onIntelligent Information Technology, Beijing, China2002,455–459
[70] Z.-H. Zhou and M.-L. Zhang. Solving multi-instance problems withclassifier ensemble based on constructive clustering[J]. Knowledge andInformation Systems2007,11(2):155–170
[71] Zhang M-L, Zhou Z-H. Adapting RBF neural networks to multi-instancelearning[J]. Neural Process Letter2006,23(1):1–26
[72] Y. Chen and J. Z. Wang. Image Categorization by Learning andReasoning with Regions[J]. Journal of Machine Learning Research2004,5:913-939
[73] Y. Chen, J. Bi, and J. Z. Wang. MILES: Multiple-instance learning viaembedded instance selection[J]. IEEE Transactions on Pattern Analysisand Machine Intelligence2006,28(12):1931–1947
[74] Maron O, Ratan A L. Multiple-instance learning for natural sceneclassification[C]. In: Proceedings of the15th International Conference onMachine Learning, Madison, WI,1998,341-349
[75] C. Yang and T. Lozano-Perez. Image Database Retrieval withMultiple-Instance Learning Techniques[C]. Proc. InternationalConference on Data Engineering,2000,233-243
[76] Zhou Z-H, Jiang K, Li M. Multi-instance learning based web mining[J].Applied Intelligence2005,22(2):135–147
[77] Liang Zhu, Bo Zhao, Yang Gao. Multi-class Multi-instance Learning forLung Cancer Image Classification Based on Bag Feature Selection[J].Fifth International Conference on Fuzzy Systems and KnowledgeDiscovery, FSKD2008,2:487–492
[78] Weidmann N, Frank E, Pfahringer B. A two-level learning method forgeneralized multi-instance problems[C]. In: Proceedings of the14thEuropean Conference on Machine Learning, Cavtat-Dubrovnik, Croatia,2003,468-479
[79] Fung, G., et al. Multiple instance learning for computer aided diagnosis.In Advances in neural information processing systems2007,19:425-432
[80] B. Krishnapuram, et al. Multiple-instance learning improves CADdetection of masses in digital mammography[C]. In Proceedings of the9th international workshop on Digital Mammography,2008,350–357
[81] D. Wu, J. Bi, K. Boyer, A min-max framework of cascaded classifier withmultiple instance learning for computer aided diagnosis[C]. in: ComputerVision and Pattern Recognition (CVPR),2009,1359–1366
[82] H.D. Cheng, J. Shan, W. Ju, Y. Guo, L. Zhang. Automated breast cancerdetection and classification using ultrasound images: a survey[J]. PatternRecognition2010,43(1):299–317
[83] Kotsiantis S: Supervised Machine Learning: A Review of ClassificationTechniques[J]. Informatica2007,31:249-268
[84] Dietterich, T.G., Jain, A., Lathrop, R. H.,&Lozano-Pérez, T. Acomparison of dynamic reposing and tangent distance for drug activityprediction[C]. Advances in Neural Information Processing Systems1994,6:216–223. San Mateo: Morgan Kaufmann
[85] Edgar, G.A. Measure, topology, and fractal geometry (3rd print)[M].1995Springer-Verlag.
[86] Cover, T., Hart, P. Nearest neighbor pattern classification[J]. IEEETransactions on Information Theory1967,13(1):21–27
[87] Atkeson C, Moore A, Schaal S. Locally weighted learning[J]. ArtificialIntelligence Review,1997,11(1):11–73
[88] Vapnik V., Bottou L. Local algorithms for pattern recognition anddependencies estimation[J]. Neural Computation.1993,5(6):893-909
[89] J.A. Noble, D. Boukerroui. Ultrasound image segmentation: a survey[J].IEEE Transactions on Medical Imaging2006,25,(8):987–1010
[90]宇传华，徐勇勇．ROC分析的基本原理[J]．中华流行病学杂志，1998,19(2-A)：413-415
[91] Fawcett, T. An introduction to ROC analysis[J]. Pattern RecognitionLetter2006,27:861874
[92] Wagner R F, Smith S W, Sandrik J M, Lopez H. Statistics of speckle inultrasound B-scans[J]. IEEE Transactions on Sonics and Ultrasonics.1983,30:156–163
[93] H.D. Cheng, X.H. Jiang, Y. Sun, J.L. Wang, Color image segmentation:advances and prospects[J]. Pattern Recognition2001,34(12):2259–2281
[94] Y.H. Guo, H.D. Cheng, J.H. Huang, J.W. Tian, W. Zhao, L.T. Sun, Y.X.Su, Breast ultrasound image enhancement using fuzzy logic[J].Ultrasound in Medicine and Biology2006,32(2):237–247
[95] Pal, S. and Dutta-Majumder, D. Fuzzy Mathematical Approach to PatternRecognition[M]. Halsted Press,1986
[96] Haralick RM, Shanmugam HK, Dinstein I. Textural features for imageclassification[J]. IEEE Trans Syst Man Cybern1973,3:610–621
[97] F.M.J. Frank, J.M. Thijssen, Characterization of echographic imagetexture by cooccurrence matrix parameters[J]. Ultrasound Med. Biol.1997,23:559–571
[98] Foulds J, Frank E: A review of multi-instance learning assumptions[J].Knowl Eng Rev2010,25(1):1–25
[99] Harris, J.R., Lippman, M.E., Morrow, M., and Osborne, C.K., Eds..Diseases of the Breast2nd Ed[M]. Philadelphia: Lippincott Willams&Wilkins2000
[100] Kohonen T. Self-organizing map[C]. Proc IEEE1990,78(9):1464–1480
[101] J. Vesanto and E. Alhoniemi. Clustering of the self-organizing map[J].IEEE Trans. Neural Networks2000,11:586–600
[102] D. Blei, A. Ng, and M. Jordan. Latent dirichlet allocation[J]. Journal ofMachine Learning Research,3:993–1022,2003
[103] Fei-Fei, L and P. Perona A Bayesian Hierarchical Model for LearningNatural Scene Categories[C]. Proc. of IEEE Computer Vision and PatternRecognition.2005,524–531
[104] K. Wagstaff, S. Rogers, and S. Schroedl, Constrained K-means clusteringwith background knowledge[C]. in Proc.8th Int. Conf. MachineLearning,2001,577–584
[105] V.Vapnik. The nature of statistical learning theory[M]. Springer, NewYork,1995
[106] Kotsiantis S: Supervised Machine Learning: A Review of ClassificationTechniques[J]. Informatica2007,31:249-268
[107]罗建文，白净.超声弹性成像的研究进展.中国医疗器械信息[J].2005,11(5)：23-31
[108]任新平，詹维伟，超声弹性成像在乳腺疾病中的应用，诊断学理论与实践[J].2007,6(2):166-168
[109] T. Shiina, et al. Clinical assessment of real-time, freehand elasticityimaging system based on the combined autocorrelation method[C]. inProc. IEEE Ultrason. Symp.,2003,664–667
[110] Garra BS, Cespedes EI, Ophir J, et al. Elastography of breast lesions:initial clinical results[J]. Radiology1997,202:79–86
[111] Itoh A, Ueno E, Tohno E, et al. Breast disease: clinical application of USelastography for diagnosis[J]. Radiology2006,239:341–350
[112] Curiel L, Souchon R, Rouvière O, Gelet A, Chapelon JY. Elastographyfor the follow-up of high-intensity focused ultrasound prostate cancertreatment: initial comparison with MRI[J]. Ultrasound in Medicine&Biology2005,31:1461–1468
[113] Miyanaga N, Akaza H, Yamakawa M, et al. Tissue elasticity imaging fordiagnosis of prostate cancer: a preliminary report[J]. International Journalof Urology2006,13:1514–1518
[114] Lyshchik A, Higashi T, Asato R, et al. Thyroid gland tumor diagnosis atUS elastography[J]. Radiology2005,237:202–211
[115] Kagoya R, Monobe H, Tojima H, et al. Utility of elastography fordifferential diagnosis of benign and malignant thyroid nodules[J].Otolaryngology-Head Neck Surgery2010,143:230–234