Based On The Characteristics Of The Vector Field And Visualization Of Medical Images

Feature based visualization has already been a very attractive and challenging part of the visualization research. Many applications are given in vector field and medical visualization. In this dissertation, we present an overview of the techniques and development of the feature based visualization and study the key algorithms in these applications.Texture based visualization for vector field, such as Line Integral Convolution (LIC), can effectively depict the global features of the vector data without losing local information. This kind of methods can afford high spatial resolution images but it is a time consuming job. To improve the performance of the algorithm, we develop a method making use of the texture hardware of the display card. We analyze the feature extraction techniques of the vector field visualization and implement a robust algorithm for topological feature detecting. Then we map the topological structures to the final images to enhance the feature displaying.The computer animation techniques can be used to explore the vector field. We discuss and compare the conventional animation algorithms and implement an algorithm mapping the phase to the magnitude of vector data based on the original line integral convolution technique. Making use of the texture hardware, we can produce the animation in real time.Next we turn our attention to mapping multi-dimensional values in vector field visualization. To display multi-dimensional information on the output image, some techniques such as color mapping, bump mapping are used to show the direction, orientation and magnitude of the vector field. We present several techniques to effectively visualize multi-dimensional values using 3D rendering, local image contrast and texture. Our goal is to enable researchers to obtain a meaningful visual summary of the vector field through providing images in which the important features of multiple scalar values can be understood.Concerned with the medical visualization, medical image segmentation is playing important role in the field of operation planning and simulation, medical education and medical research. We have studied the techniques of medical image segmentation and present an efficient adaptive region growing algorithm that estimate its parameters by studying the characteristics in the local region. The parameters are obtained by analyzing the statistical information using certain cluster algorithm whichcan reduce the interaction with users. The experiment results show the approach is reliable and efficient.Finally we study the techniques of web based visualization. With the rapid development of the Internet, Web based techniques provide a good computational and data share environment. We designed the prototype architectures of Web based visualization based on VTK and VRML and tested our system under the CFD and medical datasets.