Advertisement

Multimedia Tools and Applications

, Volume 77, Issue 17, pp 21825–21845 | Cite as

Voxelwise detection of cerebral microbleed in CADASIL patients by leaky rectified linear unit and early stopping

  • Yu-Dong Zhang
  • Xiao-Xia Hou
  • Yi Chen
  • Hong Chen
  • Ming Yang
  • Jiquan Yang
  • Shui-Hua Wang
Article

Abstract

It is important to detect cerebral microbleed voxels from the brain image of cerebral autosomal-dominant arteriopathy with subcortical infarcts and Leukoencephalopathy (CADASIL) patients. Traditional manual method suffers from intra-observe and inter-observe variability. In this study, we used the susceptibility weighted imaging (SWI) to scan 10 CADASIL patients and 10 healthy controls. We used slicing neighborhood processing (SNP) to extract “input” and “target” dataset from the 20 brain volumetric images. Afterwards, the undersampling technique was employed to handle the class-imbalanced problem. The single-hidden layer feedforward neural-network with scaled conjugate gradient was used as the classifier. We compared three activation functions: logistic sigmoid (LOSI), rectified linear unit (ReLU), and leaky rectified linear unit (LReLU). Early stopping and K-fold cross validation (CV) was used to avoid overfitting and statistical analysis. In the experiment, we generated 68,847 CMB voxels, and 68,829 non-CMB voxels. We observed that LReLU achieved the best result with a sensitivity of 93.05%, a specificity of 93.06%, and an accuracy of 93.06%. We also observed the effect of early stopping and K-fold CV. We found the optimal number of hidden neuron was 10 by grid searching method. Besides, our method performs better than three state-of-the-art methods. The results show our method is promising. In addition, LReLU is a better activation function that may replace traditional logistic sigmoid function in other applications.

Keywords

CADASIL Cerebral microbleed Magnetic resonance imaging Susceptibility weighted imaging Class-imbalanced problem Logistic sigmoid Leaky rectified linear unit 

Notes

Acknowledgments

This paper was supported by NSFC (61602250), Natural Science Foundation of Jiangsu Province (BK20150983), Leading Initiative for Excellent Young Researcher (LEADER) of Ministry of Education, Culture, Sports, Science and Technology-Japan (16809746), Program of Natural Science Research of Jiangsu Higher Education Institutions (16KJB520025), Open Research Fund of Hunan Provincial Key Laboratory of Network Investigational Technology (2016WLZC013), Open Fund of Fujian Provincial Key Laboratory of Data Intensive Computing (BD201607), Open Program of Jiangsu Key Laboratory of 3D Printing Equipment and Manufacturing (3DL201602).

References

  1. 1.
    Agarwal P (2016) Artificial intelligence and its applications 2014. Math Probl Eng. doi: 10.1155/2016/3871575
  2. 2.
    Aghdam HH et al (2015)Toward an optimal convolutional neural network for traffic sign recognition. In 8th International Conference on Machine Vision. Barcelona: Spie-Int Soc Optical Engineering. p. 98750KGoogle Scholar
  3. 3.
    Aghdam HH et al (2016) Recognizing traffic signs using a practical deep neural network. In 2nd Iberian Robotics Conference (ROBOT). Lisbon, PORTUGAL: Springer-Verlag Berlin pp 399-410Google Scholar
  4. 4.
    Ali S et al (2016) Can-CSC-GBE: developing cost-sensitive classifier with Gentleboost ensemble for breast cancer classification using protein amino acids and imbalanced data. Comput Biol Med 73:38–46CrossRefGoogle Scholar
  5. 5.
    Ando S (2016) Classifying imbalanced data in distance-based feature space. Knowl Inf Syst 46(3):707–730CrossRefGoogle Scholar
  6. 6.
    Babaie-Kafaki S, Ghanbari R (2014) A modified scaled conjugate gradient method with global convergence for nonconvex functions. Bulletin Of the Belgian Mathematical Society-Simon Stevin 21(3):465–477MathSciNetMATHGoogle Scholar
  7. 7.
    Balochian S (2014) Artificial intelligence and its applications. Math Probl Eng. doi: 10.1155/2014/840491
  8. 8.
    Barnes SRS et al (2011) Semiautomated detection of cerebral microbleeds in magnetic resonance images. Magn Reson Imaging 29(6):844–852CrossRefGoogle Scholar
  9. 9.
    Barrow E et al (2015) Deep dropout artificial neural networks for recognising digits and characters in natural images. In 22nd International Conference on Neural Information Processing (ICONIP). Istanbul: Springer Int Publishing Ag. pp. 29–37Google Scholar
  10. 10.
    Bian W et al (2013) Computer-aided detection of radiation-induced cerebral microbleeds on susceptibility-weighted MR images. Neuroimage-Clinical 2:282–290CrossRefGoogle Scholar
  11. 11.
    Borkar P et al (2016) Employing speeded scaled conjugate gradient algorithm for multiple contiguous feature vector frames: an approach for traffic density state estimation. In 1st International conference on information security & privacy. Nagpur: Elsevier Science Bv. pp. 740–747Google Scholar
  12. 12.
    Chen Y, Chen X-Q (2016) Sensorineural hearing loss detection via discrete wavelet transform and principal component analysis combined with generalized eigenvalue proximal support vector machine and Tikhonov regularization. Multimedia Tools and Applications, doi: 10.1007/s11042–016–4087-6 (Online)
  13. 13.
    Chen M et al (2015a) Detection of dendritic spines using wavelet-based conditional symmetric analysis and regularized morphological shared-weight neural networks. Comput Math Methods Med. doi: 10.1155/2015/454076
  14. 14.
    Chen M et al (2015b) AIWAC: affective interaction through wearable computing and cloud technology. IEEE Wirel Commun 22(1):20–27CrossRefGoogle Scholar
  15. 15.
    Dahl GE et al (2013) Improving deep neural networks for lvcsr using rectified linear units and dropout. In International Conference on Acoustics, Speech And Signal Processing. Vancouver: IEEE. pp. 8609–8613Google Scholar
  16. 16.
    Esmaeelzadeh SR et al (2015) Long-term streamflow forecasts by adaptive neuro-fuzzy inference system using satellite images and K-fold cross-validation (case study: Dez, Iran). KSCE J Civ Eng 19(7):2298–2306CrossRefGoogle Scholar
  17. 17.
    Famouri M et al (2015) Fast linear svm validation based on early stopping in iterative learning. International Journal Of Pattern Recognition And Artificial Intelligence 29(8):1551013MathSciNetCrossRefGoogle Scholar
  18. 18.
    Fazlollahi A et al (2015) Computer-aided detection of cerebral microbleeds in susceptibility-weighted imaging. Comput Med Imaging Graph 46:269–276CrossRefGoogle Scholar
  19. 19.
    Feng C (2015) Feed-forward neural network optimized by hybridization of PSO and ABC for abnormal brain detection. Int J Imaging Syst Technol 25(2):153–164CrossRefGoogle Scholar
  20. 20.
    Gayathri K, Kumarappan N (2015) Double circuit EHV transmission lines fault location with RBF based support vector machine and reconstructed input scaled conjugate gradient based neural network. International Journal Of Computational Intelligence Systems 8(1):95–105CrossRefGoogle Scholar
  21. 21.
    Goodwin JA et al (2015) Susceptibility-weighted phase imaging and oxygen extraction fraction measurement during sedation and sedation recovery using 7 T MRI. J Neuroimaging 25(4):575–581CrossRefGoogle Scholar
  22. 22.
    Greenberg SM et al (2009) Cerebral microbleeds: a guide to detection and interpretation. Lancet Neurol 8(2):165–174CrossRefGoogle Scholar
  23. 23.
    Gregoire SM et al (2009) The microbleed anatomical rating scale (MARS) reliability of a tool to map brain microbleeds. Neurology 73(21):1759–1766CrossRefGoogle Scholar
  24. 24.
    Hara K et al (2015) Analysis of function of rectified linear unit used in deep learning. In International Joint Conference on Neural Networks. Killarney: IEEE. pp. 144–151Google Scholar
  25. 25.
    Imaizumi T et al (2007) Dynamics of dot-like hemosiderin spots on T2*-weighted MRIs associated with stroke recurrence. J Neuroimaging 17(3):204–210CrossRefGoogle Scholar
  26. 26.
    Ji G (2014) Fruit classification using computer vision and feedforward neural network. J Food Eng 143:167–177CrossRefGoogle Scholar
  27. 27.
    Klein K et al (2016) A bayesian modelling approach with balancing informative prior for analysing imbalanced data. Plos One 11(4):e0152700CrossRefGoogle Scholar
  28. 28.
    Klimstra M, Zehr EP (2008) A sigmoid function is the best fit for the ascending limb of the Hoffmann reflex recruitment curve. Exp Brain Res 186(1):93–105CrossRefGoogle Scholar
  29. 29.
    Kolus A et al (2015) Adaptive neuro-fuzzy inference systems with k-fold cross-validation for energy expenditure predictions based on heart rate. Appl Ergon 50:68–78CrossRefGoogle Scholar
  30. 30.
    Kuijf HJ et al (2012) Efficient detection of cerebral microbleeds on 7.0 T MR images using the radial symmetry transform. NeuroImage 59(3):2266–2273CrossRefGoogle Scholar
  31. 31.
    Lee JB Lee JH (2015) An iterative undersampling of extremely imbalanced data using csvm. in 7th international conference on machine vision. Milan, ITALY: Spie-Int Soc Optical Engineering. pp. 3460–3471Google Scholar
  32. 32.
    Lee Y et al (2014) A new susceptibility-weighted image reconstruction method for the reduction of background phase artifacts. Magn Reson Med 71(3):1324–1335CrossRefGoogle Scholar
  33. 33.
    Levati E et al (2016) Moonlighting transcriptional activation function of a fungal sulfur metabolism enzyme. Sci Report 6:25165CrossRefGoogle Scholar
  34. 34.
    Li B, Rong XW (2013) Review and performance analysis of single hidden layer sequential learning algorithms of feed-forward neural networks. In 25th chinese control and decision conference. Guiyang, Peoples R China: IEEE. pp. 2170–2175Google Scholar
  35. 35.
    Liao B et al (2015) An image retrieval method for binary images based on DBN and softmax classifier. IETE Tech Rev 32(4):294–303CrossRefGoogle Scholar
  36. 36.
    Liu A (2015) Magnetic resonance brain image classification via stationary wavelet transform and generalized eigenvalue proximal support vector machine. Journal of Medical Imaging and Health Informatics 5(7):1395–1403CrossRefGoogle Scholar
  37. 37.
    Lu HM (2016) Facial emotion recognition based on biorthogonal wavelet entropy, fuzzy support vector machine, and stratified cross validation. IEEE Access 4:8375–8385CrossRefGoogle Scholar
  38. 38.
    Murray V et al (2010) Multiscale AM-FM demodulation and image reconstruction methods with improved accuracy. IEEE Trans Image Process 19(5):1138–1152MathSciNetCrossRefMATHGoogle Scholar
  39. 39.
    Njikam ANS, Zhao H (2016) A novel activation function for multilayer feed-forward neural networks. Appl Intell 45(1):75–82CrossRefGoogle Scholar
  40. 40.
    Osoba O, Kosko B (2016) The noisy expectation-maximization algorithm for multiplicative noise injection. Fluctuation And Noise Letters 15(1):23 Article ID: 1650007CrossRefGoogle Scholar
  41. 41.
    Peng B et al (2016) Image processing methods to elucidate spatial characteristics of retinal microglia after optic nerve transection. Scientific Reports 6:21816CrossRefGoogle Scholar
  42. 42.
    Roy S et al (2015) Cerebral microbleed segmentation from susceptibility weighted images. Proc SPIE 9413. doi: 10.1117/12.2082237
  43. 43.
    Rutten JW et al (2015) The NOTCH3 score: a pre-clinical CADASIL biomarker in a novel human genomic NOTCH3 transgenic mouse model with early progressive vascular NOTCH3 accumulation. Acta Neuropathol Commun 3:89CrossRefGoogle Scholar
  44. 44.
    Seghier ML et al (2011) Microbleed detection using automated segmentation (midas): a new method applicable to standard clinical mr images. Plos One 6(3):e17547CrossRefGoogle Scholar
  45. 45.
    Sepasian M et al. (2008) Image enhancement for fingerprint minutiae-based algorithms using clahe, standard deviation analysis and sliding neighborhood. In world congress on engineering and computer science. San Francisco: Int Assoc Engineers-IAENG. pp. 1199–1203Google Scholar
  46. 46.
    Shah P et al (2016) Inverse scattering using a joint L1-L2 norm-based regularization. IEEE Trans Antennas Propag 64(4):1373–1384MathSciNetCrossRefMATHGoogle Scholar
  47. 47.
    Shao Y et al (2011) Comparison of early stopping criteria for neural-network-based subpixel classification. IEEE Geosci Remote Sens Lett 8(1):113–117CrossRefGoogle Scholar
  48. 48.
    Tan RYY, Markus HS (2016) CADASIL: migraine, encephalopathy, stroke and their inter-relationships. Plos One 11(6):e0157613CrossRefGoogle Scholar
  49. 49.
    Tsivgoulis G et al (2016) Risk of symptomatic intracerebral hemorrhage after intravenous thrombolysis in patients with acute ischemic stroke and high cerebral microbleed burden a meta-analysis. JAMA Neurology 73(6):675–683CrossRefGoogle Scholar
  50. 50.
    Valmoria MS et al (2014) Clopedogril loading in acute ischemic stroke patients with cerebral microbleed does not increase the risk of hemorrhage. Ann Neurol 76:S88–S89Google Scholar
  51. 51.
    Vitali P et al (2014) Cerebral microbleed causing an acute stroke-like episode in a CADASIL patient. Can J Neurol Sci 41(5):661–663CrossRefGoogle Scholar
  52. 52.
    Wei L (2015) Fruit classification by wavelet-entropy and feedforward neural network trained by fitness-scaled chaotic ABC and biogeography-based optimization. Entropy 17(8):5711–5728Google Scholar
  53. 53.
    Wesolowski W et al (2015) Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) - literature review apropos an autopsy case. Pol J Pathol 66(3):323–329CrossRefGoogle Scholar
  54. 54.
    Wu L (2011) A hybrid method for MRI brain image classification. Expert Syst Appl 38(8):10049–10053CrossRefGoogle Scholar
  55. 55.
    Wu J (2016) Fruit classification by biogeography-based optimization and feedforward neural network. Expert Syst 33(3):239–253CrossRefGoogle Scholar
  56. 56.
    Wu XX Liu JG (2009) A new early stopping algorithm for improving neural network generalization. In 2nd International Conference on Intelligent Computation Technology And Automation. Changsha, Peoples R China: IEEE Computer Soc pp 15–18Google Scholar
  57. 57.
    Wu YW et al (2014) Image denoising with rectified linear units. In 21st International Conference on Neural Information Processing (ICONIP). Kuching: Springer-Verlag Berlin. pp. 142-149Google Scholar
  58. 58.
    Yang M (2016) Dual-tree complex wavelet transform and twin support vector machine for pathological brain detection. Appl Sci 6(6). doi: 10.3390/app6060169
  59. 59.
    Zadeh MR et al (2010) Daily outflow prediction by multi layer perceptron with logistic sigmoid and tangent sigmoid activation functions. Water Resour Manag 24(11):2673–2688CrossRefGoogle Scholar
  60. 60.
    Zainuddin Z, Fard SP (2015) Approximation of multivariate 2 pi-periodic functions by multiple 2 pi-periodic approximate identity neural networks based on the universal approximation theorems. in 11th International Conference on Natural Computation (ICNC). Zhangjiajie, Peoples R China: IEEE pp 8–13Google Scholar
  61. 61.
    Zhan T (2016) Pathological brain detection by artificial intelligence in magnetic resonance imaging scanning. Prog Electromagn Res 156:105–133CrossRefGoogle Scholar
  62. 62.
    Zhan TM, Chen Y (2016) Multiple sclerosis detection based on biorthogonal wavelet transform, RBF kernel principal component analysis, and logistic regression. IEEE Access 4:7567–7576MathSciNetCrossRefGoogle Scholar
  63. 63.
    Zhang Y (2016) GroRec: a Group-centric Intelligent Recommender System Integrating Social, Mobile and Big Data Technologies. IEEE Trans Serv Comput:1–15Google Scholar
  64. 64.
    Zhang Y et al (2014) CAP: community activity prediction based on big data analysis. IEEE Netw 28(4):52–57MathSciNetCrossRefGoogle Scholar
  65. 65.
    Zhang Y et al (2015a) Health-CPS: healthcare cyber-physical system assisted by cloud and big data. IEEE Systems Journal 99:1–8Google Scholar
  66. 66.
    Zhang Y et al (2015b) CADRE: cloud-assisted drug recommendation service for online pharmacies. Mobile Networks & Applications 20(3):348–355CrossRefGoogle Scholar
  67. 67.
    Zhou X et al (2015) Detection of pathological brain in MRI scanning based on wavelet-entropy and naive Bayes classifier. In Bioinformatics and biomedical engineering. Granada, Spain: Springer International Publishing. pp. 201–209Google Scholar

Copyright information

© Springer Science+Business Media New York 2017

Authors and Affiliations

  • Yu-Dong Zhang
    • 1
    • 2
    • 3
  • Xiao-Xia Hou
    • 3
  • Yi Chen
    • 1
    • 2
  • Hong Chen
    • 3
  • Ming Yang
    • 4
  • Jiquan Yang
    • 1
  • Shui-Hua Wang
    • 1
    • 5
  1. 1.School of Computer Science and TechnologyNanjing Normal UniversityNanjingChina
  2. 2.Hunan Provincial Key Laboratory of Network Investigational TechnologyChangshaChina
  3. 3.Department of NeurologyFirst Affiliated Hospital of Nanjing Medical UniversityNanjingChina
  4. 4.Department of Radiology, Nanjing Children’s HospitalNanjing Medical UniversityNanjingChina
  5. 5.Department of Electrical Engineering, The City College of New YorkCUNYNew YorkUSA

Personalised recommendations