Research On Multi-Transfer Function Technology And 3D Medical Image Visualization System Design And Implement

Visualization on medical images is a hot spot field. It is an important application of computer graphics and image processing in biomedicine engineering. Visualization of medical images is widely used in diagnostic, surgery planning and simulating. Direct Volume Rendering is one of the most important techniques for visualizing medical volume data.Transfer function specification is arguably the most important task in volume visualization that will highlight those aspects of the volume data that are of interest to the user. As the traditional transfer function is difficult to extract the complex feature, this thesis implements a multidimensional transfer function base on the scalar value and the gradient magnitude. For this method is involved the problem of the noise impact and fixed size of the gradient calculation, a transfer function base local data features design method is presented in this thesis ,by examining relationships among three eigenvalues of inertia matrix.In this thesis, the two materials that form the boundary are labeled L and H. We design and implement a transfer function based on the LH histogram. The experiments tell that the LH transfer function is easy to select the boundaries while arches could be overlap in scalar value and gradient magnitude histogram. In mirrored LH space both sides of boundary can be classified independently.This thesis introduce two local transfer functions, their fields are users interactive selected in three sections or rendering space. They can easily isolate the regions of interesting while uninteresting regions contain the same range of data value. Finally, a medical image volume rendering system, which is based on texture-based hardware accelerated volume rendering technique, is designed and implemented with excellent result.