Color perception of diffusion tensor images using hierarchical manifold learning
作者：
Highlights：
• A new method is proposed for color perception of DTI using hierarchical manifold learning.
• The novel method has better regional structures, which enhance image feature information and improve the visual effect.
• Our method reduces the computational complexity of the classical method [9] from O(N3) to O(N2).
• A new metric approximately estimates the geodesic distance between the seeds to avoid both great error and double counting.
• The proposed method focuses on the color perception of medical images application.
摘要
•A new method is proposed for color perception of DTI using hierarchical manifold learning.•The novel method has better regional structures, which enhance image feature information and improve the visual effect.•Our method reduces the computational complexity of the classical method [9] from O(N3) to O(N2).•A new metric approximately estimates the geodesic distance between the seeds to avoid both great error and double counting.•The proposed method focuses on the color perception of medical images application.
论文关键词：Diffusion tensor images,High dimensional data,Nonlinear dimensionality reduction,Color perception,Algebraic multigrid,Multi-scale graph partitioning
论文评审过程：Received 30 January 2016, Revised 27 August 2016, Accepted 21 September 2016, Available online 21 September 2016, Version of Record 27 November 2016.
论文官网地址：https://doi.org/10.1016/j.patcog.2016.09.021
