Deep Learning-Based Automatic Segmentation of the Proximal Femur from MR Images
This chapter addresses the problem of segmentation of proximal femur in 3D MR images. We propose a deeply supervised 3D U-net-like fully convolutional network for segmentation of proximal femur in 3D MR images. After training, our network can directly map a whole volumetric data to its volume-wise labels. Inspired by previous work, multi-level deep supervision is designed to alleviate the potential gradient vanishing problem during training. It is also used together with partial transfer learning to boost the training efficiency when only small set of labeled training data are available. The present method was validated on 20 3D MR images of femoroacetabular impingement patients. The experimental results demonstrate the efficacy of the present method.
KeywordsMRI Segmentation Femoroacetabular impingement (FAI) Proximal femur Deep learning Fully Convolutional Network (FCN) Deep supervision
This chapter was modified from the paper published by our group in the MICCAI 2017 Workshop on Machine Learning in Medical Imaging (Zeng and Zheng, MLMI@MICCAI 2017: 274-282). The related contents were reused with the permission. This study was partially supported by the Swiss National Science Foundation via project 205321_163224/1.
- 10.Krizhevsky A, ISutskever, Hinton G (2012) Imagenet classification with deep convolutional neural networks. In: Pereira F, Burges CJC, Bottou L, Weinberger KQ (eds) Advances in neural information processing systems, vol 25. Curran Associates, Inc., Red Hook, pp 1097–1105Google Scholar
- 11.Long J, Shelhamer E, Darrell T (2015) Fully convolutional networks for semantic segmentation. In: Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR 2015), pp 3431–3440, BostonGoogle Scholar
- 12.Prasson A, Igel C, Petersen K et al (2013) Deep feature learning for knee cartilage segmentation using a triplanar convolutional neural network. In: Proceedings of the 16th international conference on medical image computing and computer assisted intervention (MICCAI 2013), vol 16(Pt 2), pp 246–53, NagoyaGoogle Scholar
- 13.Cicek O, Abdulkadir A, Lienkamp S, Brox T, Ronneberger O (2016) 3D u-net: learning dense volumetric segmentation from sparse annotation. In: Proceedings of the 16th international conference on medical image computing and computer assisted intervention (MICCAI 2016). LNCS, vol 9901, pp 424–432, AthensGoogle Scholar
- 14.Milletari F, Navab N, Ahmadi SA (2016) V-net: fully convolutional neural networks for volumetric medical image segmentation. In: Proceedings of the 2016 international conference on 3D vision (3DV). IEEE, pp 565–571, StanfordGoogle Scholar
- 16.Ioffe S, Szegedy C (2015) Batch normalization: accelerating deep network training by reducing internal covariate shift. In: Proceedings of international conference on machine learning (ICML 2015), LilleGoogle Scholar
- 17.Yosinski J, Clune J, Bengio Y, Lipson H (2014) How transferable are features in deep neural networks? In: Advances in neural information processing systems, pp 3320–3328, Curran Associates, Inc.Google Scholar
- 18.Deng J, Dong W, Socher R, Li LJ, Li K, Fei-Fei L (2009) ImageNet: a large-scale hierarchical image database. In: Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR 2009), Miami BeachGoogle Scholar
- 19.Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv:1409.1556Google Scholar
- 20.Szegedy C, Liu W, Jia Y et al (2015) Going deeper with convolutions. In: Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR 2015), Boston. IEEE, pp 1–9Google Scholar
- 21.Tran D, Bourdev L, Fergus R, Torresani L, Paluri M (2015) Learning spatiotemporal features with 3D convolutional networks. In: Proceedings of the IEEE international conference on computer vision (CVPR 2015), pp 4489–4497, BostonGoogle Scholar