6 March 2018 Research on segmentation based on multi-atlas in brain MR image
Author Affiliations +
Proceedings Volume 10610, MIPPR 2017: Parallel Processing of Images and Optimization Techniques; and Medical Imaging; 106100D (2018) https://doi.org/10.1117/12.2288033
Event: Tenth International Symposium on Multispectral Image Processing and Pattern Recognition (MIPPR2017), 2017, Xiangyang, China
Abstract
Accurate segmentation of specific tissues in brain MR image can be effectively achieved with the multi-atlas-based segmentation method, and the accuracy mainly depends on the image registration accuracy and fusion scheme. This paper proposes an automatic segmentation method based on the multi-atlas for brain MR image. Firstly, to improve the registration accuracy in the area to be segmented, we employ a target-oriented image registration method for the refinement. Then In the label fusion, we proposed a new algorithm to detect the abnormal sparse patch and simultaneously abandon the corresponding abnormal sparse coefficients, this method is made based on the remaining sparse coefficients combined with the multipoint label estimator strategy. The performance of the proposed method was compared with those of the nonlocal patch-based label fusion method (Nonlocal-PBM), the sparse patch-based label fusion method (Sparse-PBM) and majority voting method (MV). Based on our experimental results, the proposed method is efficient in the brain MR images segmentation compared with MV, Nonlocal-PBM, and Sparse-PBM methods.
© (2018) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Yuejing Qian, "Research on segmentation based on multi-atlas in brain MR image", Proc. SPIE 10610, MIPPR 2017: Parallel Processing of Images and Optimization Techniques; and Medical Imaging, 106100D (6 March 2018); doi: 10.1117/12.2288033; https://doi.org/10.1117/12.2288033
PROCEEDINGS
8 PAGES


SHARE
Back to Top