Multi-scale Volumetric ConvNet with Nested Residual Connections for Segmentation of Anterior Cranial Base

Anterior cranial base (ACB) is known as the growth-stable structure. Automatic segmentation of the ACB is a prerequisite to superimpose orthodontic inter-treatment cone-beam computed tomography (CBCT) images. The automatic ACB segmentation is still a challenging task because of the ambiguous intensity distributions around fine-grained structures and artifacts due to the limited radiation dose. We propose a fully automatic segmentation of the ACB from CBCT images by a volumetric convolutional network with nested residual connections (NRN). The multi-scale feature fusion in the NRN not only promotes the information flows, but also introduces the supervision to multiple intermediate layers to speed up the convergence. The multi-level shortcut connections augment the feature maps in the decompression pathway and the end-to-end voxel-wise label prediction. The proposed NRN has been applied to the ACB segmentation from clinically-captured CBCT images. The quantitative assessment over the practitioner-annotated ground truths demonstrates the proposed method produces improvements to the state-of-the-arts.

[1]  James L Ackerman,et al.  Cranial base superimposition for 3-dimensional evaluation of soft-tissue changes. , 2010, American journal of orthodontics and dentofacial orthopedics : official publication of the American Association of Orthodontists, its constituent societies, and the American Board of Orthodontics.

[2]  Patrick van der Smagt,et al.  CNN-based Segmentation of Medical Imaging Data , 2017, ArXiv.

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

[4]  Carlos Flores-Mir,et al.  Anterior cranial-base time-related changes: A systematic review. , 2014, American journal of orthodontics and dentofacial orthopedics : official publication of the American Association of Orthodontists, its constituent societies, and the American Board of Orthodontics.

[5]  Seunghoon Hong,et al.  Learning Deconvolution Network for Semantic Segmentation , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).

[6]  Hao Chen,et al.  Volumetric ConvNets with Mixed Residual Connections for Automated Prostate Segmentation from 3D MR Images , 2017, AAAI.

[7]  Chunming Li,et al.  Level set evolution without re-initialization: a new variational formulation , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[8]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[9]  D. Louis Collins,et al.  Patch-based segmentation using expert priors: Application to hippocampus and ventricle segmentation , 2011, NeuroImage.

[10]  Lisa Tang,et al.  Deep 3D Convolutional Encoder Networks With Shortcuts for Multiscale Feature Integration Applied to Multiple Sclerosis Lesion Segmentation , 2016, IEEE Transactions on Medical Imaging.

[11]  Junzhou Huang,et al.  Deformable Segmentation via Sparse Shape Representation , 2011, MICCAI.

[12]  Thomas Brox,et al.  3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation , 2016, MICCAI.

[13]  Zhuowen Tu,et al.  Holistically-Nested Edge Detection , 2015, ICCV.

[14]  Jitendra Malik,et al.  Hypercolumns for object segmentation and fine-grained localization , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[15]  Gregory Shakhnarovich,et al.  Feedforward semantic segmentation with zoom-out features , 2014, CVPR.

[16]  Seyed-Ahmad Ahmadi,et al.  V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation , 2016, 2016 Fourth International Conference on 3D Vision (3DV).