Study Of Visualization Techniques Based On Diffusion Tensor Imaging

As a relatively new MRI technique, diffusion tensor imaging (DTI) makes it possible to explore the white matter of brain in vivo, quantitatively and non-invasively. As a non-invasive method for analysis of brain internal structure, DTI plays an important role. in the field of neurophysiology, neuroanatomy and neurosurgery.DTI bases on anisotropic diffusion properties of water molecules. As we known, even if we see a bowl of water looks static with naked eye, in fact the internal water molecules constantly collide randomly with each other. This kind of motion is named as "Brownian motion". Diffusion of water molecules along the fiber axis is faster than along vertical orientation, this phenomenon called anisotropy; In homogeneous medium, diffusion rate is the same of all orientation, this phenomenon called isotropy. The geometrical and physical properties of white matter determine the diffusion rate and orientation, so the human tissues can show different diffusion properties. It's the basic principle of diffusion tensor imaging.Different from traditional gray-scalar and color medical image data, DTI image data is a complex tensor field data. It is a three dimensional volume, composed of a series of two-dimensional slice, and each voxel corresponds to a 2nd-order symmetric positive definite tensor. How to dig the meaningful information in tensor data and seek real and intuitive representation to display these results has become one of the major mission of DTI research.The article is study of commonly used visualization techniques based on DTI, including colorcoding, tensor glyphs and isosurface.Colorcoding techniques include colorcoding anisotropy indices and colorcoding main diffusion direction. The six values for the different directions that are measured by DTI are reduced to one anisotropy indices. Scalar values can be colorcode(gray) with a lookuptable. A result of the paper is colorcoding Fraction anisotropy(FA). FA indicates white matter, it is hightest inside the thick regions of uniformly anisotropyic diffusion, such as inside the corpus callosum. Using the main diffusion direction is another colorcoding method. The main eigenvector of the diffusion is determined RGB components. Red is left to right, Green is anterior to posterior, Bule is feet to head.Tensor glyphs can show the information of diffusion tensor. The paper gives three methods for conveying all tensor diffusion attributes in a restricted set of discrete locations, by ellipsoidal glyphs, cubical glyphs and superquadric glyphs. The length of the axis is determined by size of the eigenvalues and the direction of axes of the tensor glyphs are determined by the eigenvector. Be compared with colorcoding, tensor orientation and shape can display diffusion attributes more vividly.Isosurface method for DTI visualization extracts Fractional anisotropy as opacity and color. Opacity is 0.0 to 1.0 depending on whether FA is below or above the indicated threshold. The paper uses Marching Cubes and Ray Casting algorithms to visualize basic 3D structure. Compare to tensor glyphs, isosurface can show continuous information of white matter.