Multi-scale Networks for Segmentation of Brain Magnetic Resonance Images

Measuring the distribution of major brain tissues using magnetic resonance (MR) images has attracted extensive research efforts. Due to its remarkable success, deep learning-based image segmentation has been applied to this problem, in which the size of patches usually represents a tradeoff between complexity and accuracy. In this paper, we propose the multi-size-and-position neural network (MSPNN) for brain MR image segmentation. Our contributions include (1) jointly using U-Nets trained on large patches and back propagation neural networks (BPNNs) trained on small patches for segmentation, and (2) adopting the convolutional auto-encoder (CAE) to restore MR images before applying them to BPNNs. We have evaluated this algorithm against five widely used brain MR image segmentation approaches on both synthetic and real MR studies. Our results indicate that the proposed algorithm can segment brain MR images effectively and provide precise distribution of major brain tissues.

[1]  David Dagan Feng,et al.  Hidden Markov random field model based brain MR image segmentation using clonal selection algorithm and Markov chain Monte Carlo method , 2014, Biomed. Signal Process. Control..

[2]  Yaozong Gao,et al.  Fully convolutional networks for multi-modality isointense infant brain image segmentation , 2016, 2016 IEEE 13th International Symposium on Biomedical Imaging (ISBI).

[3]  D.G. Tzikas,et al.  The variational approximation for Bayesian inference , 2008, IEEE Signal Processing Magazine.

[4]  D. Louis Collins,et al.  Design and construction of a realistic digital brain phantom , 1998, IEEE Transactions on Medical Imaging.

[5]  Hao Chen,et al.  VoxResNet: Deep voxelwise residual networks for brain segmentation from 3D MR images , 2017, NeuroImage.

[6]  Qiang Chen,et al.  Fuzzy Local Gaussian Mixture Model for Brain MR Image Segmentation , 2012, IEEE Transactions on Information Technology in Biomedicine.

[7]  E. Halgren,et al.  Dynamic Statistical Parametric Mapping Combining fMRI and MEG for High-Resolution Imaging of Cortical Activity , 2000, Neuron.

[8]  Max A. Viergever,et al.  Automatic Segmentation of MR Brain Images With a Convolutional Neural Network , 2016, IEEE Transactions on Medical Imaging.

[9]  Shuiwang Ji,et al.  Deep convolutional neural networks for multi-modality isointense infant brain image segmentation , 2015, NeuroImage.

[10]  A. D'Amico,et al.  Evaluation of three-dimensional finite element-based deformable registration of pre- and intraoperative prostate imaging. , 2001, Medical physics.

[11]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[12]  Assas Ouarda,et al.  Improvement of MR brain images segmentation based on interval type-2 fuzzy C-Means , 2015, 2015 Third World Conference on Complex Systems (WCCS).

[13]  Herng-Hua Chang,et al.  A level set based deformable model for segmentation of human brain MR images , 2014, 2014 7th International Conference on Biomedical Engineering and Informatics.

[14]  Torsten Rohlfing,et al.  Image Similarity and Tissue Overlaps as Surrogates for Image Registration Accuracy: Widely Used but Unreliable , 2012, IEEE Transactions on Medical Imaging.

[15]  Ulla Ruotsalainen,et al.  Genetic Algorithms for Finite Mixture Model Based Voxel Classification in Neuroimaging , 2007, IEEE Transactions on Medical Imaging.

[16]  Giovanni Montana,et al.  Deep neural networks for anatomical brain segmentation , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[17]  James L. McClelland,et al.  Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations , 1986 .

[18]  Yogita K. Dubey,et al.  Segmentation of brain MR images using rough set based intuitionistic fuzzy clustering , 2016 .

[19]  Jürgen Schmidhuber,et al.  Stacked Convolutional Auto-Encoders for Hierarchical Feature Extraction , 2011, ICANN.

[20]  Mark W. Woolrich,et al.  Advances in functional and structural MR image analysis and implementation as FSL , 2004, NeuroImage.