Medical Image Three Dimensional Visualization Based On Direct Volume Rendering

According to the configuration cells, three dimensional medical image visualization methods can be divided into two categories, one is surface rendering methods based on micro-surfaces, and another is volume rendering methods based on voxels. Surface rendering algorithms use micro-surfaces to fit the objects' surface, the results of which just show the outer surfaces of tissues, can not display the inner structures in deep levels; Direct Volume Rendering(DVR) algorithms use voxels to construct objects, and the intact three dimensional information of human tissues can be reserved, but due to the difficulty of setting Opacity Transformer Function(OTF). which is caused by voxels masking off each other, that makes DVR rendering results usually only show the outer structures of tissues. For the three dimensional visualization of ultrasound images, the inherent properties in ultrasound imaging, such as the speckle noise, the lower gray dynamic range, quite wide blurry area. make the ultrasound three dimensional reconstruction very difficult, and all of the traditional medical three dimensional visualization technologies need preprocessing steps, otherwise it is difficult to obtain useful rendering results.In this thesis, we improve Levoy's DVR algorithm to widen its use in medical multi-dimensional image visualization. For MRI data set visualization, we use sample piecewise liner function to construct OTF, and treated tissues as a kind of...