Visual Exploration of Diffierences Among DTI Fiber Models. Issue 5 (1st September 2018)

• Record Type: Journal Article
• Authors: Mei, Honghui; Guo, Fangzhou; Chen, Haidong; Chen, Yi; Guo, Fangzhou
• Is Part Of: Chinese journal of electronics. Volume 27:Issue 5(2018)
• Page Start: 959
• Page End: 967
• Publication Date: 2018-09-01
• Subjects: Diffiusion tensor imaging -- Fiber tracking -- Diffierence visualization -- Visual exploration; biodiffusion -- biomedical MRI -- data visualisation -- feature extraction -- medical image processing; visual exploration -- DTI fiber models -- noninvasive tools -- fiber tracking -- Diffusion tensor imaging datasets -- complex 3D fiber model -- interactive exploration -- DTI images -- signal noises -- tracking method -- atlas construction -- uncertainty analysis -- 3D space -- visual clutters -- two‐phase projection technique -- 2D map -- 2D maps; Electronics -- Periodicals; Electronics -- China -- Periodicals; Electronics; China; Periodicals; 621.38105
• DOI: 10.1049/cje.2018.06.015
• Languages: English
• ISSNs: 1022-4653

Abstract

Abstract : In‐vivo studies of fibrous structures require non‐invasive tools, of which one is fiber tracking based on Diffiusion tensor imaging (DTI) datasets. Different fiber models can be produced from diffierent DTI images, which may vary from subject to subject due to variations in anatomy, motions in scanning, and signal noises. Additionally, parameters of the tracking method also have a great influence on resulting models. Illustrating, exploring, and analyzing diffierences among DTI fiber models are crucial for the purposes of group comparison, atlas construction, and uncertainty analysis. Conventional approaches illustrate fiber models in 3D space and explore diffierences either voxel‐wisely or fiber‐based. However, these approaches rely on accurate alignment processes and may easily be disturbed by visual clutters. We introduce a two‐phase projection technique to illustrate a complex 3D fiber model with a unique 2D map to characterize features for further exploration and analysis. Moreover, regions of significant diffierences among the maps are marked out. In these 2D maps, diffierences can be easily distinguished without occlusions that often occur in 3D spaces. To facilitate comparative analysis from multiple perspectives, we design an interface for interactive exploration. The effiectiveness of our approach is evaluated with two datasets.