188宝金博页面版

Amira - a Highly Interactive Systemfor Visual Data AnalysisDetlev Stalling, Malte Westerhoff, and Hans-Christian HegeZuse Institute Berlin (ZIB), Germany1IntroductionWhat characteristics should a good visualization system hold? What kinds of data should itsupport? What capabilities should it provide? Of course, the answers depend on the particulartask and application. For some users a visualization system may be nothing more than a simpleimage viewer or plotting program. For others it is integrated software dedicated to their personalfield of work, such as a computer algebra program or a finite-element simulation system. Whilein such integrated systems visualization is usually just an add-on, there are also many specializedsystems whose primary focus is upon visualization itself.On the one hand, there are many self-contained special-purpose programs written for partic-ular applications. Examples include flow visualization systems, finite-element post-processors,and volume rendering software for medical images. On the other hand, several general-purposevisualization systems have been developed since scientific visualization became an independentfield ofresearch in the late 1980s. These systems are not targeted to a particular application area,but provide many different modules which can be combined in numerous ways, often adheringto the data-flow principle and providing means for ‘visual programming’ [23, 4, 17, 1, 5, 15, 7].In these ways, custom pipelines can be built to solve specific visualization problems. Al-though these visualization environments are very flexible and powerful, they are usually moredifficult to use than special-purpose software. In addition, a major drawback induced by thedata-flow principle or pipelining approach is the lack of sophisticated user interaction. Any op-eration which requires manual interaction, such as segmenting a medical image into differentregions, editing a polygonal surface model or molecular structures, or cropping and selectingdifferent parts of a complex finite-element model, is difficult to incorporate into a pipeline ofmodules that is executed automatically. One may argue that these interactive tasks are not visu-alization problems per se. But it is a matter of fact that such operations require visual supportand are essential for solving problems in many application areas.In order to close the gap between the ease of use, power, and interactivity of monolithicspecial-purpose software, and the flexibility and extensibility of data-flow oriented visualizationenvironments, the software system amira has been designed. Initially developed by the scientificvisualization group at the Zuse Institute Berlin (ZIB), today amira is available as a commercialproduct together with several extensions and add-ons [2]. One major focus of the software isthe visualization and analysis of volumetric data which is common in medicine, biology, andmicroscopy. These volumes can be displayed and segmented, 3D polygonal models can be re-constructed, and these models can be further processed and converted into tetrahedral volume1