Home Friction Article
PDF (2.4 MB)
Cite
EndNote(RIS) BibTeX
Collect
Submit Manuscript
Show Outline
Figures (5)

Tables (2)
Table 1
Table 2
Research Article | Open Access

Morphological residual convolutional neural network (M-RCNN) for intelligent recognition of wear particles from artificial joints

Xiaobin HU1, Jian SONG2(), Zhenhua LIAO3, Yuhong LIU4(), Jian GAO5, Bjoern MENZE1, Weiqiang LIU3,4
Department of Computer Science, Technical University of Munich, Garching 85748, Germany
School of Biomedical Engineering, Sun Yat-sen University, Guangzhou 510006, China
Key Laboratory of Biomedical Materials and Implant Devices, Research Institute of Tsinghua University in Shenzhen, Shenzhen 518057, China
State Key Laboratory of Tribology, Tsinghua University, Beijing 100084, China
Mckelvey School of Engineering, Washington University in Saint Louis, St. Louis, MO 63130, USA
Show Author Information

Abstract

Finding the correct category of wear particles is important to understand the tribological behavior. However, manual identification is tedious and time-consuming. We here propose an automatic morphological residual convolutional neural network (M-RCNN), exploiting the residual knowledge and morphological priors between various particle types. We also employ data augmentation to prevent performance deterioration caused by the extremely imbalanced problem of class distribution. Experimental results indicate that our morphological priors are distinguishable and beneficial to largely boosting overall performance. M-RCNN demonstrates a much higher accuracy (0.940) than the deep residual network (0.845) and support vector machine (0.821). This work provides an effective solution for automatically identifying wear particles and can be a powerful tool to further analyze the failure mechanisms of artificial joints.

Keywords

wear particles classifiermorphological priorsdata augmentationdeep residual network

Electronic Supplementary Material

Download File(s)
101237TH-2022-4-560_ESM.pdf (296.6 KB)

References

[1]
Mattei L, Di Puccio F, Piccigallo B, Ciulli E. Lubrication and wear modelling of artificial hip joints: A review. Tribol Int 44(5):532-549 (2011)
Crossref Google Scholar
[2]
Carr B C, Goswami T. Knee implants-Review of models and biomechanics. Mater Des 30(2):398-413 (2009)
Crossref Google Scholar
[3]
Song J, Xiang D D, Wang S, Liao Z H, Lu J Z, Liu Y H, Liu W Q, Peng Z X. In vitro wear study of PEEK and CFRPEEK against UHMWPE for artificial cervical disc application. Tribol Int 122:218-227 (2018)
Crossref Google Scholar
[4]
Meng Y G, Xu J, Jin Z M, Prakash B, Hu Y Z. A review of recent advances in tribology. Friction 8(2):221-300 (2020)
Crossref Google Scholar
[5]
Sawano H, Warisawa S, Ishihara S. Study on long life of artificial joints by investigating optimal sliding surface geometry for improvement in wear resistance. Precision Eng 33(4):492-498 (2009)
Crossref Google Scholar
[6]
Bao Q B, Yuan H A. Artificial disc technology. Neurosurg Focus 9(4):e14 (2000)
Crossref Google Scholar
[7]
Harris W H. Osteolysis and particle disease in hip replacement: A review. Acta Orthop Scand 65(1):113-123 (1994)
Crossref Google Scholar
[8]
Gallo J, Goodman S B, Konttinen Y T, Raska M. Particle disease: Biologic mechanisms of periprosthetic osteolysis in total hip arthroplasty. Innate Immun 19(2):213-224 (2013)
Crossref Google Scholar
[9]
Sukur E, Akman Y E, Ozturkmen Y, Kucukdurmaz F. Particle disease: A current review of the biological mechanisms in periprosthetic osteolysis after hip arthroplasty. Open Orthop J 10:241-251 (2016)
Crossref Google Scholar
[10]
Nine M J, Choudhury D, Hee A C, Mootanah R, Osman N A A. Wear debris characterization and corresponding biological response: Artificial hip and knee joints. Materials 7(2):980-1016 (2014)
Crossref Google Scholar
[11]
McMullin B T, Leung M Y, Shanbhag A S, McNulty D, Mabrey J D, Agrawal C M. Correlating subjective and objective descriptors of ultra high molecular weight wear particles from total joint prostheses. Biomaterials 27(5):752-757 (2006)
Crossref Google Scholar
[12]
Liu H T, Ge S R, Gao S F, Wang S B. Comparison of wear debris generated from ultra high molecular weight polyethylene in vivo and in artificial joint simulator. Wear 271(5-6):647-652 (2011)
Crossref Google Scholar
[13]
Song J, Chen F F, Liu Y H, Wang S, He X, Liao Z H, Mu X H, Yang M Y, Liu W Q, Peng Z X. Insight into the wear particles of PEEK and CFRPEEK against UHMWPE for artificial cervical disc application: Morphology and immunoreaction. Tribol Int 144:106093 (2020)
Crossref Google Scholar
[14]
Myshkin N, Kovalev A. Adhesion and surface forces in polymer tribology—A review. Friction 6(2):143-155 (2018)
Crossref Google Scholar
[15]
Khan M, Kuiper J H, Richardson J B. Can cobalt levels estimate in-vivo wear of metal-on-metal bearings used in hip arthroplasty? Proc Inst Mech Eng H 221(8):929-942 (2007)
Crossref Google Scholar
[16]
Peng Y P, Cai J H, Wu T H, Cao G Z, Kwok N, Peng Z X. WP-DRnet: A novel wear particle detection and recognition network for automatic ferrograph image analysis. Tribol Int 151:106379 (2020)
Crossref Google Scholar
[17]
Wang S, Wu T H, Shao T, Peng Z X. Integrated model of BP neural network and CNN algorithm for automatic wear debris classification. Wear 426-427:1761-1770 (2019)
Crossref Google Scholar
[18]
Laghari M S, Boujarwah A. Wear particle texture classification using artificial neural networks. Int J Pattern Recognit Artif Intell 13(3):415-428 (1999)
Crossref Google Scholar
[19]
Kumar M, Mukherjee P S, Misra N M. Advancement and current status of wear debris analysis for machine condition monitoring: A review. Ind Lubricat Tribol 65(1):3-11 (2013)
Crossref Google Scholar
[20]
Yan X P, Xu X J, Sheng C X, Yuan C Q, Li Z X. Intelligent wear mode identification system for marine diesel engines based on multi-level belief rule base methodology. Meas Sci Technol 29(1):015110 (2018)
Crossref Google Scholar
[21]
Stachowiak G P, Podsiadlo P, Stachowiak G W. Shape and texture features in the automated classification of adhesive and abrasive wear particles. Tribol Lett 24(1):15-26 (2006)
Crossref Google Scholar
[22]
Peng Y P, Cai J H, Wu T H, Cao G Z, Kwok N, Zhou S X, Peng Z X. A hybrid convolutional neural network for intelligent wear particle classification. Tribol Int 138:166-173 (2019)
Crossref Google Scholar
[23]
Eckold D G, Dearn K D, Shepherd D E T. The evolution of polymer wear debris from total disc arthroplasty. Biotribology 1-2:42-50 (2015)
Crossref Google Scholar
[24]
Guo H X, Li Y J, Shang J, Gu M Y, Huang Y Y, Gong B. Learning from class-imbalanced data: Review of methods and applications. Expert Syst Appl 73:220-239 (2017)
Crossref Google Scholar
[25]
Kingma D, Ba J. Adam: A method for stochastic optimization. arXiv:1412.6980 (2015)
[26]
Seltzer M L, Yu D, Wang Y Q. An investigation of deep neural networks for noise robust speech recognition. In Proceedings of 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, Vancouver, Canada, 2013: 7398-7402.
Crossref Google Scholar
[27]
Kim I, Baek W, Kim S. Spatially attentive output layer for image classification. In Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seattle, USA, 2020: 9533-9542.
Crossref Google Scholar
[28]
Zhao W Z, Du S H. Spectral-spatial feature extraction for hyperspectral image classification: A dimension reduction and deep learning approach. IEEE Trans Geosci Remote Sens 54(8):4544-4554 (2016)
Crossref Google Scholar
[29]
Sbai O, Couprie C, Aubry M. Impact of base dataset design on few-shot image classification. In Proceedings of the 16th European Conference on Computer Vision, Glasgow, UK, 2020: 597-613.
Crossref Google Scholar
[30]
Ioffe S, Szegedy C. Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167 (2015)
Google Scholar
[31]
Maas A L, Hannun A Y, Ng A Y. Rectifier nonlinearities improve neural network acoustic models. In Proceedings of the 30th International Conference on Machine Learning, Atlanta, Georgia, USA, 2013: 3.
Google Scholar
[32]
Goodfellow I, Bengio Y, Courville A. Deep Learning. Cambridge (USA): MIT Press, 2016.
[33]
Chen Y S, Jiang H L, Li C Y, Jia X P, Ghamisi P. Deep feature extraction and classification of hyperspectral images based on convolutional neural networks. IEEE Trans Geosci Remote Sens 54(10):6232-6251 (2016)
Crossref Google Scholar
[34]
Zou K H, O’Malley A J, Mauri L. Receiver-operating characteristic analysis for evaluating diagnostic tests and predictive models. Circulation 115(5):654-657 (2007)
Crossref Google Scholar
Friction
Volume 10 Issue 4,
April 2022
Pages 560-572
DOI: 10.1007/s40544-021-0516-2
Cite this article:
HU X, SONG J, LIAO Z, et al. Morphological residual convolutional neural network (M-RCNN) for intelligent recognition of wear particles from artificial joints. Friction, 2022, 10(4): 560-572. https://doi.org/10.1007/s40544-021-0516-2
Metrics & Citations  
Article History
Copyright
Rights and Permissions
Return

10.1007/s40544-021-0516-2.T001The precision and accuracy values of wear particle recognition using different methods: support vector machine (SVM), deep residual network (ResNet), deep residual network with data augmentation (ResNet+Aug), and morphological residual convolutional neural network (M-RCNN). A, B, C, D, and E donate the flake-like, spherical, aggregated, rod-like, and zonal particles, respectively.

Wear particleSVMResNetResNet+AugM-RCNN
PrecisionA0.5840.6320.6560.855
BNANNANNAN0.714
C0.2500.5450.6670.821
D00.5000.6150.857
E000.3330.933
Overall0.5530.6120.6490.851
AccuracyA0.5590.6170.6700.872
B0.9630.9630.9630.979
C0.7870.8350.8400.915
D0.9150.9260.9410.979
E0.8830.8830.8830.957
Overall0.8210.8450.8600.940

10.1007/s40544-021-0516-2.T002The ablation study of wear particle recognition using our different schemes: with/without morphological priors and with/without ensemble mechanism: M-RCNN without morphological priors, M-RCNN and ensembled models. A, B, C, D, and E donate the flake-like, spherical, aggregated, rod-like, and zonal particles, respectively.

Wear particleM-RCNN without morphological priorsM-RCNNEnsembled
PrecisionA0.8110.8550.856
B0.5710.7140.714
C0.8060.8210.852
D0.8180.8570.857
E0.7500.9330.933
Overall0.7980.8510.856
AccuracyA0.8240.8720.878
B0.9680.9790.979
C0.9200.9150.920
D0.9630.9790.979
E0.9200.9570.957
Overall0.9190.9400.943