Tensor decomposition processes for interpolation of diffusion magnetic resonance imaging

作者：

Highlights：

• A Novel probabilistic framework for interpolation of dMRI-HOT is proposed.

• The framework is based on Gaussian processes combined with tensor decompositions.

• Results demonstrate improvements in accuracy and generalization to any rank.

• Validation on synthetic and real data evaluating anisotropy levels and fiber tracts.

摘要

•A Novel probabilistic framework for interpolation of dMRI-HOT is proposed.•The framework is based on Gaussian processes combined with tensor decompositions.•Results demonstrate improvements in accuracy and generalization to any rank.•Validation on synthetic and real data evaluating anisotropy levels and fiber tracts.

论文关键词：Diffusion magnetic resonance imaging,Higher order tensors,Interpolation,Probabilistic models,Tensor decomposition

论文评审过程：Received 29 November 2017, Revised 8 September 2018, Accepted 3 October 2018, Available online 4 October 2018, Version of Record 10 October 2018.

论文官网地址：https://doi.org/10.1016/j.eswa.2018.10.005