Slice-wise optimization algorithm for diffusion tensor estimation

Wei Liu, Zhenyu Zhou, Xiaozheng Liu, Xu Yan, Guang Yang, Zunliang Wang, Yongdi Zhou, S. Peterson Bradley, Dongrong Xu

Research output: Contribution to journalArticlepeer-review


A slice-wise optimization algorithm for diffusion tensor estimation is proposed to improve the accuracy of estimating tensors by using diffusion weighted imaging (DWI) data which contain artifacts. Firstly, three types of common artifacts (wavelike, motion-between-slice and contrast artifacts) are qualitatively analyzed in DWI data. Three types of features (wavelet signature, similarity and correlation) are extracted to identify these three artifacts, respectively. Thus, a slice with or without artifacts can be distinguished. Then, tensors can be reconstructed by using the slices tagged without any artifacts. The simulation results show that the three features are effective to identify related artifacts in DWI data. A high discrimination capability (>90%) can be achieved for identifying wavelike and motion-between-slice artifacts. The experimental results using real datasets demonstrate that, compared with other similar algorithms, the proposed algorithm can improve the bias found in color-encoded fractional anisotropy map more effectively, and can provide more accurate directionality information to analyze white matter structure.

Original languageEnglish (US)
Pages (from-to)30-34
Number of pages5
JournalDongnan Daxue Xuebao (Ziran Kexue Ban)/Journal of Southeast University (Natural Science Edition)
Issue number1
StatePublished - Jan 2013
Externally publishedYes


  • Diffusion tensor imaging
  • Diffusion weighted imaging
  • Gray-level co-occurrence matrix
  • Mutual information
  • Tensor optimization
  • Wavelet signature

ASJC Scopus subject areas

  • Engineering(all)


Dive into the research topics of 'Slice-wise optimization algorithm for diffusion tensor estimation'. Together they form a unique fingerprint.

Cite this