基于深度学习特征的乳腺肿瘤分类模型评估
|
摘要
目的本文结合深度学习特征(DF)和传统图像特征(HCF)特点,利用多分类器融合的方法建立一个乳腺肿瘤分类模型,并深入评估和分析不同深度学习网络特征的肿瘤分类性能。方法回顾性分析106例乳腺肿瘤患者的头尾位和内外倾斜位投影的全数字乳腺成像数据。首先从肿瘤区域提取23维HCF(12维形态及11维纹理特征),用t检验进行显著性特征选择;然后分别从3个卷积神经网络模型提取不同维度DF,在实验中,3个不同深度学习网络产生了相应DF,分别是AlexNet,VGG16和GoogLeNet;最后结合2个投影数据的DF和HCF,采用多分类器的融合模型对特征进行训练和测试,实验重点分析不同DF在肿瘤分类上的性能。结果结合DF和HCF建立的分类模型比使用单独HCF的分类模型表现出更好的性能;相比于其它网络框架,DF_(AlexNet)和HCF的结合表现出更高精度的分类结果。结论结合DF和HCF的特征方法建立一个分类模型,对于良恶性乳腺肿瘤具有优秀的鉴别能力,且泛化能力更强,能作为临床辅助诊断工具。
Objective To develop a deep features-based model to classify benign and malignant breast lesions on full-filed digital mammography. Method The data of full-filed digital mammography in both craniocaudal view and mediolateral oblique view from 106 patients with breast neoplasms were analyzed. Twenty-three handcrafted features(HCF) were extracted from the images of the breast tumors and a suitable feature set of HCF was selected using t-test. The deep features(DF) were extracted from the 3 pre-trained deep learning models, namely AlexNet, VGG16 and GoogLeNet. With abundant breast tumor information from the craniocaudal view and mediolateral oblique view, we combined the two extracted features(DF and HCF)as the two-view features. A multi-classifier model was finally constructed based on the combined HCF and DF sets. The classification ability of different deep learning networks was evaluated. Results Quantitative evaluation results showed that the proposed HCF + DF model outperformed HCF model, and AlexNet produced the best performances among the 3 deep learning models. Conclusion The proposed model that combines DF and HCF sets of breast tumors can effectively distinguish benign and malignant breast lesions on full-filed digital mammography.
Objective To develop a deep features-based model to classify benign and malignant breast lesions on full-filed digital mammography. Method The data of full-filed digital mammography in both craniocaudal view and mediolateral oblique view from 106 patients with breast neoplasms were analyzed. Twenty-three handcrafted features(HCF) were extracted from the images of the breast tumors and a suitable feature set of HCF was selected using t-test. The deep features(DF) were extracted from the 3 pre-trained deep learning models, namely AlexNet, VGG16 and GoogLeNet. With abundant breast tumor information from the craniocaudal view and mediolateral oblique view, we combined the two extracted features(DF and HCF)as the two-view features. A multi-classifier model was finally constructed based on the combined HCF and DF sets. The classification ability of different deep learning networks was evaluated. Results Quantitative evaluation results showed that the proposed HCF + DF model outperformed HCF model, and AlexNet produced the best performances among the 3 deep learning models. Conclusion The proposed model that combines DF and HCF sets of breast tumors can effectively distinguish benign and malignant breast lesions on full-filed digital mammography.
引文
[1]Siegel RL,Miller KD,Jemal A.Cancer statistics,2018[J].CACancer J Clin,2018,68(1):7-30.
[2]Lehman CD,Arao RF,Sprague BL,et al.National performance benchmarks for modern screening digital mammography:update from the breast cancer surveillance consortium[J].Radiology,2017,283(1):49-58.
[3]Coldman A,Phillips N,Wilson C,et al.Pan-canadian study of mammography screening and mortality from breast cancer[J].J Natl Cancer Inst,2014,106(11):1-7.
[4]Kumar V,Gu Y,Basu S,et al.Radiomics:the process and the challenges[J].Magn Reson Imaging,2012,30(9):1234-48.
[5]Xie WY,Li YS,Ma YD.Breast mass classification in digital mammography based on extreme learning machine[J].Neurocomputing,2016,173(p3):930-41.
[6]Muramatsu C,Hara T,Endo T,et al.Breast mass classification on mammograms using radial local ternary patterns[J].Comput Biol Med,2016,72(1):43-53.
[7]Velikova M,Samulski M,Lucas PJ.Improved mammographic CADperformance using multi-view information:a bayesian network framework[J].PhysMed Biol,2009,54(5):1131-47.
[8]Tan MX,Pu JT,Zheng B.Reduction of false-positive recalls using a computerized mammographic image feature analysis scheme[J].PhysMed Biol,2014,59(15):4357-73.
[9]Schmidhuber J.Deep learning in neural networks:an overview[J].Neural Networks,2015,61(1):85-117.
[10]He KM,Zhang XY,Ren SQ,et al.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition(CPVR),2016:770-8.
[11]Spanhol FA,Oliveira LS,Cavalin PR,et al.Deep features for breast cancer histopathological image classification[C]//2017 IEEEInternational Conference on Systems,Man and Cybemetics,2017:1868-73.
[12]Dhungel N,Carneiro G,Bradley AP.The automated learning of deep features for breast mass classification from mammograms[C]//Lecture Notes in Computer Science,2016:106-14.
[13]Jiao ZC,Gao XB,Wang Y,et al.A deep feature based framework for breast masses classification[J].Neurocomputing,2016,197(2):221-31.
[14]Wang X,Albregtsen F,Foyn B.Texture features from gray level gap length matrix[C]//Iapr Workshop on Machine Vision Applications,2009:375-8.
[15]Krizhevsky A,Sutskever I,Hinton G.ImageNet classification with deep convolutional neural networks[C]//International Conference on Neural Information Processing Systems,2012:1097-105.
[16]Simonyan K,Zisserman A.Very deep convolutional networks for large-scale image recognition[J].Comput Sci,2014,23(14):1409-556.
[17]Szegedy C,LiuW,JiaYQ,et al.Going deeper with convolutions[C]//2015 IEEE Conference on Computer Vision and Pattern Recognition(CVPR),2015:1-9.
[18]Shen D,Wu G,Suk HI.Deep learning in medical image analysis[J].Annu Rev Biomed Eng,2017,19(1):221-48.
[19]Shin HC,Roth HR,Gao M,et al.Deep convolutional neural networks for computer-aided detection:CNN architectures,dataset characteristics and transfer learning[J].IEEE Trans Med Imaging,2016,35(5):1285-98.
[20]Srikant R,Agrawal R.Mining generalized association rules[C]//International Conference onVery Large Data Bases,1995:407-19.
[21]Breiman L.Random forest[J].Mach Learn,2001,45(1):5-32.
[22]Conklin JD.Applied logistic regression[J].Technometrics,2013,44(1):81-2.
[23]Zou WY,Ng AY,Zhu S,et al.Deep learning of invariant features via simulated fixations in video[J].Adv Neural Inf Process Syst,2012,25(12):3212-20.
[24]Hilden J,Glasziou P.Regret graphs,diagnostic uncertainty and Youden's Index[J].StatMed,2015,15(10):969-86.
[25]Li H,Mendel KR,Lee JH,et al.Deep learning in breast cancer risk assessment:evaluation of fine-tuned convolutional neural networks on a clinical dataset of FFDMs[C]//Medical Imanging 2018:Computer-aided Diagnosis,2018:10575-8.
[26]Zhou B,Garcia AL,Xiao J,et al.Learning deep features for scene recognition using places database[C]//International Conference on Neural Information Processing Systems:mit press,2014:358-63.
[27]Zhang Q,Xiao Y,Dai W,et al.Deep learning based classification of breast tumors with shear-wave elastography[J].Ultrasonics,2016,72(2):150-7.
[28]Kuncheva L.Combining pattern classifiers:methods and algorithms[M].NewYork:Wiley-Interscience,2004.
[29]Marcus G.Deep learning:a critical appraisal[J].CoRR,2018,476(13):1-26.
[30]Kayla M,Hui L,Deepa S,et al.Transfer learning from convolutional neural networks for computer-aided diagnosis:a comparison of digital breast tomosynthesis and full-field digital mammography[J].Acad Radiol,2018,25(10):1-9.
[31]周志华.机器学习[M].北京:清华大学出版社,2016.
[2]Lehman CD,Arao RF,Sprague BL,et al.National performance benchmarks for modern screening digital mammography:update from the breast cancer surveillance consortium[J].Radiology,2017,283(1):49-58.
[3]Coldman A,Phillips N,Wilson C,et al.Pan-canadian study of mammography screening and mortality from breast cancer[J].J Natl Cancer Inst,2014,106(11):1-7.
[4]Kumar V,Gu Y,Basu S,et al.Radiomics:the process and the challenges[J].Magn Reson Imaging,2012,30(9):1234-48.
[5]Xie WY,Li YS,Ma YD.Breast mass classification in digital mammography based on extreme learning machine[J].Neurocomputing,2016,173(p3):930-41.
[6]Muramatsu C,Hara T,Endo T,et al.Breast mass classification on mammograms using radial local ternary patterns[J].Comput Biol Med,2016,72(1):43-53.
[7]Velikova M,Samulski M,Lucas PJ.Improved mammographic CADperformance using multi-view information:a bayesian network framework[J].PhysMed Biol,2009,54(5):1131-47.
[8]Tan MX,Pu JT,Zheng B.Reduction of false-positive recalls using a computerized mammographic image feature analysis scheme[J].PhysMed Biol,2014,59(15):4357-73.
[9]Schmidhuber J.Deep learning in neural networks:an overview[J].Neural Networks,2015,61(1):85-117.
[10]He KM,Zhang XY,Ren SQ,et al.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition(CPVR),2016:770-8.
[11]Spanhol FA,Oliveira LS,Cavalin PR,et al.Deep features for breast cancer histopathological image classification[C]//2017 IEEEInternational Conference on Systems,Man and Cybemetics,2017:1868-73.
[12]Dhungel N,Carneiro G,Bradley AP.The automated learning of deep features for breast mass classification from mammograms[C]//Lecture Notes in Computer Science,2016:106-14.
[13]Jiao ZC,Gao XB,Wang Y,et al.A deep feature based framework for breast masses classification[J].Neurocomputing,2016,197(2):221-31.
[14]Wang X,Albregtsen F,Foyn B.Texture features from gray level gap length matrix[C]//Iapr Workshop on Machine Vision Applications,2009:375-8.
[15]Krizhevsky A,Sutskever I,Hinton G.ImageNet classification with deep convolutional neural networks[C]//International Conference on Neural Information Processing Systems,2012:1097-105.
[16]Simonyan K,Zisserman A.Very deep convolutional networks for large-scale image recognition[J].Comput Sci,2014,23(14):1409-556.
[17]Szegedy C,LiuW,JiaYQ,et al.Going deeper with convolutions[C]//2015 IEEE Conference on Computer Vision and Pattern Recognition(CVPR),2015:1-9.
[18]Shen D,Wu G,Suk HI.Deep learning in medical image analysis[J].Annu Rev Biomed Eng,2017,19(1):221-48.
[19]Shin HC,Roth HR,Gao M,et al.Deep convolutional neural networks for computer-aided detection:CNN architectures,dataset characteristics and transfer learning[J].IEEE Trans Med Imaging,2016,35(5):1285-98.
[20]Srikant R,Agrawal R.Mining generalized association rules[C]//International Conference onVery Large Data Bases,1995:407-19.
[21]Breiman L.Random forest[J].Mach Learn,2001,45(1):5-32.
[22]Conklin JD.Applied logistic regression[J].Technometrics,2013,44(1):81-2.
[23]Zou WY,Ng AY,Zhu S,et al.Deep learning of invariant features via simulated fixations in video[J].Adv Neural Inf Process Syst,2012,25(12):3212-20.
[24]Hilden J,Glasziou P.Regret graphs,diagnostic uncertainty and Youden's Index[J].StatMed,2015,15(10):969-86.
[25]Li H,Mendel KR,Lee JH,et al.Deep learning in breast cancer risk assessment:evaluation of fine-tuned convolutional neural networks on a clinical dataset of FFDMs[C]//Medical Imanging 2018:Computer-aided Diagnosis,2018:10575-8.
[26]Zhou B,Garcia AL,Xiao J,et al.Learning deep features for scene recognition using places database[C]//International Conference on Neural Information Processing Systems:mit press,2014:358-63.
[27]Zhang Q,Xiao Y,Dai W,et al.Deep learning based classification of breast tumors with shear-wave elastography[J].Ultrasonics,2016,72(2):150-7.
[28]Kuncheva L.Combining pattern classifiers:methods and algorithms[M].NewYork:Wiley-Interscience,2004.
[29]Marcus G.Deep learning:a critical appraisal[J].CoRR,2018,476(13):1-26.
[30]Kayla M,Hui L,Deepa S,et al.Transfer learning from convolutional neural networks for computer-aided diagnosis:a comparison of digital breast tomosynthesis and full-field digital mammography[J].Acad Radiol,2018,25(10):1-9.
[31]周志华.机器学习[M].北京:清华大学出版社,2016.