Submit Manuscript  

Article Details


3D Cascaded Convolutional Networks for Multi-Vertebrae Segmentation

Author(s):

Liu Xia, Liu Xiao, Gan Quan and Wang Bo   Pages 1 - 9 ( 9 )

Abstract:


Background: Automatic approach to vertebrae segmentation from computed tomography (CT) images is very important in clinical applications. As the intricate appearance and variable archi-tecture of vertebrae across the population, cognate constructions in close vicinity, pathology, and the interconnection between vertebrae and ribs´╝îit is a challenge to propose a 3D automatic vertebrae CT image segmentation method.

Objective: The purpose of this study was to propose an automatic multi-vertebrae segmentation method for spinal CT images.

Methods: Firstly, CLAHE-Threshold-Expansion was preprocessed to improve image quality and reduce input voxel points. Then, 3D coarse segmentation fully convolutional network and cascaded finely segmentation convolutional neural network were used to complete multi-vertebrae seg-mentation and classification.

Results: The results of this paper were compared with other methods on the same datasets. Ex-perimental results demonstrated that the Dice similarity coefficient (DSC) in this paper is 94.84%, higher than the V-net and 3D U-net.

Conclusion: Method of this paper has certain advantages in automatically and accurately seg-menting vertebrae regions of CT images. Due to the easy acquisition of spine CT images. It was proved to be more conducive to clinical application of treatment that using our segmentation model to obtain vertebrae regions, combining with the subsequent 3D reconstruction and printing work.

Keywords:

CT Image, 3D Vertebra Segmentation, FCN, CNN

Affiliation:

School of automation, Harbin University of Science and Technology, Harbin 150001, School of automation, Harbin University of Science and Technology, Harbin 150001, School of automation, Harbin University of Science and Technology, Harbin 150001, School of automation, Harbin University of Science and Technology, Harbin 150001



Read Full-Text article