Advertisement

Multimedia Tools and Applications

, Volume 78, Issue 10, pp 12939–12960 | Cite as

Multi-scale CNN based on region proposals for efficient breast abnormality recognition

  • Ibtissam BakkouriEmail author
  • Karim Afdel
Article

Abstract

Mammographic pattern recognition is one of the most essential tasks in breast cancer diagnosis, and has been studied for several years now to make it suitable and faster. In this paper, we developed a novel deep Convolutional Neural Network (CNN) approach to discriminate normal from abnormal breast tissues using Gaussian pyramid representation for multi-scale analysis (Pyramid-CNN). In order to improve image processing time, we extracted representative region proposals from each mammogram using determinant of the Hessian operator. To improve performance of our model and avoid overfitting, data augmentation techniques based on geometric transformation and sub-histogram equalization were applied on all regions to increase the variance of significant mammographic samples. We evaluated our methodology on the publicly available mammography dataset such as Breast Cancer Digital Repository (BCDR) database. In comparison with the current state-of-the-art methods, the experiments show that our proposed system provides efficient results, achieving the average accuracy of 96.84%, sensitivity of 92.12%, specificity of 98.02%, precision of 92.15%, F1-score of 92.12%, and area under the receiver operating characteristic curve (AUC) of 96.76%. Hence, the study demonstrates that our proposed approach has the potential to significantly improve the conventional recognition and classification strategies for use in advanced clinical application and practice or in general, biomedical imaging field.

Keywords

Mammogram Pattern recognition Convolutional neural network (CNN) Multi-scale analysis 

Notes

Acknowledgements

The authors would like expressing their gratitude to the Department of Radiology at Hospital São João Porto, Portugal for providing the JPG images of the BCDR database which was used in this research.

References

  1. 1.
    Abeyratne U, Kinouchi Y, Oki H, Okada J, Shichijo F, Matsumoto K (1991) Artificial neural networks for source localization in the human brain. Brain Topogr 4:3–21.  https://doi.org/10.1007/bf01129661 CrossRefGoogle Scholar
  2. 2.
    Aminikhanghahi S, Shin S, Wang W, Jeon S, Son S (2016) A new fuzzy Gaussian mixture model (FGMM) based algorithm for mammography tumor image classification. Multimed Tools Appl 76:10191–10205.  https://doi.org/10.1007/s11042-016-3605-x CrossRefGoogle Scholar
  3. 3.
    Aslan F, Yuce B, Oztas Z, Ates H (2017) Evaluation of the learning curve of non-penetrating glaucoma surgery. International Ophthalmology.  https://doi.org/10.1007/s10792-017-0691-3
  4. 4.
    Bakkouri I, Afdel K (2017) Breast tumor classification based on deep convolutional neural networks. In: International Conference on advanced technologies for signal and image processing (ATSIP).  https://doi.org/10.1109/atsip.2017.8075562
  5. 5.
    Bodai B, Tuso P (2015) breast cancer survivorship: a comprehensive review of long-term medical issues and lifestyle recommendations. The Permanente Journal.  https://doi.org/10.7812/tpp/14-241
  6. 6.
    Burt P, Adelson E (1983) The Laplacian pyramid as a compact image code. IEEE Trans Commun 31:532–540.  https://doi.org/10.1109/tcom.1983.1095851 CrossRefGoogle Scholar
  7. 7.
    Carney P, Elmore J, Abraham L, Gerrity M, Hendrick R, Taplin S, Barlow W, Cutter G, Poplack S, D’Orsi C (2004) Radiologist uncertainty and the interpretation of screening. Med Decis Making 24:255–264.  https://doi.org/10.1177/0272989x04265480 CrossRefGoogle Scholar
  8. 8.
    Chen W, Samuelson F (2014) The average receiver operating characteristic curve in multireader multicase imaging studies. Br J Radiol 87:20140016.  https://doi.org/10.1259/bjr.20140016 CrossRefGoogle Scholar
  9. 9.
    Chen S, McMullan G, Faruqi A, Murshudov G, Short J, Scheres S, Henderson R (2013) High-resolution noise substitution to measure overfitting and validate resolution in 3D structure determination by single particle electron cryomicroscopy. Ultramicroscopy 135:24–35.  https://doi.org/10.1016/j.ultramic.2013.06.004 CrossRefGoogle Scholar
  10. 10.
    Choromanska A, Henaff M, Mathieu M, Ben Arous G, LeCun Y (2015) The loss surfaces of multilayer networks. Int Conf Artif Intell Statist 38:192–204Google Scholar
  11. 11.
    Coates A, Huval B, Wang T, Wu D, Catanzaro BY, Ng A (2013) Deep learning with COTS HPC systems. International Conference on Machine LearningGoogle Scholar
  12. 12.
    de Oliveira Martins L, da Silva E, Silva A, de Paiva A, Gattass M (2007) Classification of breast masses in mammogram images using Ripley’s K function and support vector machine. Mach Learn Data Min Pattern Recogn, 784–794Google Scholar
  13. 13.
    Dhungel N, Carneiro G, Bradley A (2015) Automated mass detection in mammograms using cascaded deep learning and random forests. In: International conference on digital image computing, techniques and applications (DICTA).  https://doi.org/10.1109/dicta.2015.7371234
  14. 14.
    Diz J, Marreiros G, Freitas A (2016) Applying data mining techniques to improve breast cancer diagnosis. J Med Syst,  https://doi.org/10.1007/s10916-016-0561-y
  15. 15.
    DobruchSobczak K (2012) The differentiation of the character of solid lesions in the breast in the compression sonoelastography. Part I: the diagnostic value of the ultrasound B-mode imaging in the differentiation diagnostics of solid, focal lesions in the breast in relation to the pathomorphological verification. J Ultrasonograph 12:402–419.  https://doi.org/10.15557/jou.2012.0029 CrossRefGoogle Scholar
  16. 16.
    Doi K (2007) Computer-aided diagnosis in medical imaging: historical review, current status and future potential. Comput Med Imaging Graph 31:198–211.  https://doi.org/10.1016/j.compmedimag.2007.02.002 CrossRefGoogle Scholar
  17. 17.
    Gong Y, Wang L, Guo R, Lazebnik S (2014) Multi-scale orderless pooling of deep convolutional activation features. Comput Vis – ECCV 2014:392–407Google Scholar
  18. 18.
    Hassaballah M, Abdelmgeid A, Alshazly H (2016) Image features detection, description and matching. Image Feature Detectors Descrip, 11–45Google Scholar
  19. 19.
    Hepsag P, Ozel S, Yazici A (2017) Using deep learning for mammography classification. In: International Conference on computer science and engineering (UBMK),  https://doi.org/10.1109/ubmk.2017.8093429
  20. 20.
    Hinton G, Osindero S, Teh Y (2006) A fast learning algorithm for deep belief nets. Neural Comput 18:1527–1554.  https://doi.org/10.1162/neco.2006.18.7.1527 MathSciNetCrossRefzbMATHGoogle Scholar
  21. 21.
    Jia Y, Shelhamer E, Donahue J, Karayev S, Long J, Girshick R, Guadarrama S, Darrell T (2014) Caffe: convolutional architecture for fast feature embedding. In: Proceedings of the ACM international conference on multimedia - MM ’14.  https://doi.org/10.1145/2647868.2654889
  22. 22.
    Kallenberg M, Petersen K, Nielsen M, Ng A, Diao P, Igel C, Vachon C, Holland K, Winkel R, Karssemeijer N, Lillholm M (2016) Unsupervised deep learning applied to breast density segmentation and mammographic risk scoring. IEEE Trans Med Imag 35:1322–1331.  https://doi.org/10.1109/tmi.2016.2532122 CrossRefGoogle Scholar
  23. 23.
    Kavukcuoglu K, Sermanet P, Boureau Y, Gregor K, Mathieu M, LeCun Y (2010) Learning convolutional feature hierarchies for visual recognition. Int Conf Neural Inf Process Syst 1:1090–1098Google Scholar
  24. 24.
    Keller B, Oustimov A, Wang Y, Chen J, Acciavatti R, Zheng Y, Ray S, Gee J, Maidment A, Kontos D (2015) Parenchymal texture analysis in digital mammography: robust texture feature identification and equivalence across devices. J Med Imag 2:024501.  https://doi.org/10.1117/1.jmi.2.2.024501 CrossRefGoogle Scholar
  25. 25.
    Kumar I, Bhadauria H, Virmani J, Thakur S (2017) A hybrid hierarchical framework for classification of breast density using digitized film screen mammograms. Multimed Tools Appl 76:18789–18813.  https://doi.org/10.1007/s11042-016-4340-z CrossRefGoogle Scholar
  26. 26.
    Lan Z, Lin M, Li X, Hauptmann A, Raj B (2015) Beyond Gaussian pyramid: multi-skip feature stacking for action recognition. In: IEEE Conference on computer vision and pattern recognition (CVPR),  https://doi.org/10.1109/cvpr.2015.7298616
  27. 27.
    Laroum S, Tessier D, Duval B, Hao J (2010) A local search appproach for transmembrane segment and signal peptide discrimination. Evol Comput Mach Learn Data Min Bioinform, 134–145Google Scholar
  28. 28.
    Laserson J (2011) From neural networks to deep learning. XRDS: crossroads. ACM Mag Students 18:29.  https://doi.org/10.1145/2000775.2000787 MathSciNetGoogle Scholar
  29. 29.
    LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521:436–444.  https://doi.org/10.1038/nature14539 CrossRefGoogle Scholar
  30. 30.
    Lee J, Jun S, Cho Y, Lee H, Kim G, Seo J, Kim N (2017) Deep learning in medical imaging: general overview. Korean J Radiol 18:570.  https://doi.org/10.3348/kjr.2017.18.4.570 CrossRefGoogle Scholar
  31. 31.
    Lévy D, Jain A (2016) Breast mass classification from mammograms using deep convolutional neural networks. NIPS 2016 ML4HC Workshop. arXiv:http://arXiv.org/abs/1612.00542
  32. 32.
    Li J, Niu C, Fan M (2012) Multi-scale convolutional neural networks for natural scene license plate detection. Adv Neural Netw – ISNN 2012:110–119CrossRefGoogle Scholar
  33. 33.
    López M, Moura D, Pollán R, Valiente F, Ortega C, Herrero G, Loureiro J, Fernandes T, de Araújo B (2012) BCDR: a breast cancer digital repository. In: 15th International conference on experimental mechanicsGoogle Scholar
  34. 34.
    Malar E, Kandaswamy A, Gauthaam M (2013) Multiscale and multilevel wavelet analysis of mammogram using complex neural network. Swarm, Evolut Memetic Comput, 658–668Google Scholar
  35. 35.
    Moura D, Guevara López M (2013) An evaluation of image descriptors combined with clinical data for breast cancer diagnosis. Int J Comput Assist Radiol Surg 8:561–574.  https://doi.org/10.1007/s11548-013-0838-2 CrossRefGoogle Scholar
  36. 36.
    Mousa R, Munib Q, Moussa A (2005) Breast cancer diagnosis system based on wavelet analysis and fuzzy-neural. Expert Syst Appl 28:713–723.  https://doi.org/10.1016/j.eswa.2004.12.028 CrossRefGoogle Scholar
  37. 37.
    Narváez F, Romero E (2012) Breast mass classification using orthogonal moments. Breast Imag, 64–71Google Scholar
  38. 38.
    Palma G, Bloch I, Muller S (2014) Detection of masses and architectural distortions in digital breast tomosynthesis images using fuzzy and a contrario approaches. Pattern Recogn 47:2467–2480.  https://doi.org/10.1016/j.patcog.2014.01.009 CrossRefGoogle Scholar
  39. 39.
    Pérez N, Guevara M, Silva A, Ramos I, Loureiro J (2014) Improving the performance of machine learning classifiers for Breast Cancer diagnosis based on feature selection. In: Proceedings of the 2014 federated conference on computer science and information systems,  https://doi.org/10.15439/2014f249
  40. 40.
    Petrick N, Sahiner B, Armato S, Bert A, Correale L, Delsanto S, Freedman M, Fryd D, Gur D, Hadjiiski L, Huo Z, Jiang Y, Morra L, Paquerault S, Raykar V, Samuelson F, Summers R, Tourassi G, Yoshida H, Zheng B, Zhou C, Chan H (2013) Evaluation of computer-aided detection and diagnosis systemsa). Med Phys 40:087001.  https://doi.org/10.1118/1.4816310 CrossRefGoogle Scholar
  41. 41.
    Ren S, He K, Girshick R, Sun J (2017) Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Trans Pattern Anal Mach Intell 39:1137–1149.  https://doi.org/10.1109/tpami.2016.2577031 CrossRefGoogle Scholar
  42. 42.
    Rodríguez-López V, Cruz-Barbosa R (2014) On the breast mass diagnosis using Bayesian networks. Nature-Inspired Comput Mach Learn, 474–485Google Scholar
  43. 43.
    Sainath T, Vinyals O, Senior A, Sak H (2015) Convolutional, long short-term memory, fully connected deep neural networks. In: 2015 IEEE International conference on acoustics, speech and signal processing (ICASSP),  https://doi.org/10.1109/icassp.2015.7178838
  44. 44.
    Sharma S, Khanna P (2014) Computer-aided diagnosis of malignant mammograms using Zernike moments and SVM. J Digit Imaging 28:77–90.  https://doi.org/10.1007/s10278-014-9719-7 CrossRefGoogle Scholar
  45. 45.
    Sharma K, Preet B (2016) Classification of mammogram images by using CNN classifier. In: International Conference on advances in computing, communications and informatics (ICACCI).  https://doi.org/10.1109/icacci.2016.7732477
  46. 46.
    Shen D, Wu G, Suk H (2017) Deep learning in medical image analysis. Annu Rev Biomed Eng 19:221–248.  https://doi.org/10.1146/annurev-bioeng-071516-044442 CrossRefGoogle Scholar
  47. 47.
    Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R (2014) Dropout: a simple way to prevent neural networks from overfitting. J Mach Learn Res 15:1929–1958MathSciNetzbMATHGoogle Scholar
  48. 48.
    Suzuki S, Zhang X, Homma N, Ichiji K, Sugita N, Kawasumi Y, Ishibashi T, Yoshizawa M (2016) Mass detection using deep convolutional neural network for mammographic computer-aided diagnosis. In: 55th Annual Conference of the society of instrument and control engineers of japan (SICE).  https://doi.org/10.1109/sice.2016.7749265
  49. 49.
    Tsochatzidis L, Zagoris K, Arikidis N, Karahaliou A, Costaridou L, Pratikakis I (2017) Computer-aided diagnosis of mammographic masses based on a supervised content-based image retrieval approach. Pattern Recogn 71:106–117CrossRefGoogle Scholar
  50. 50.
    Wu S, Yu S, Yang Y, Xie Y (2013) Feature and contrast enhancement of mammographic image based on multiscale analysis and morphology. Comput Math Methods Med 2013:1–8.  https://doi.org/10.1155/2013/716948 MathSciNetzbMATHGoogle Scholar
  51. 51.
    Xu X (2014) Blob detection with the determinant of the Hessian. Commun Comput Inform Sci, 72–80Google Scholar
  52. 52.
    Yoon H, Han O, Hahn H (2009) Image contrast enhancement based sub-histogram equalization technique without over-equalization noise. Int J Electric Comput Eng 3 (2):189–195Google Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2018

Authors and Affiliations

  1. 1.LabSIV, Faculty of Science, Department of Computer ScienceIbn Zohr UniversityAgadirMorocco

Personalised recommendations