Viola Ivan, Feixas Miquel, Sbert Mateu, Gröller Meister Eduard
University of Bergen, Norway.
IEEE Trans Vis Comput Graph. 2006 Sep-Oct;12(5):933-40. doi: 10.1109/TVCG.2006.152.
This paper introduces a concept for automatic focusing on features within a volumetric data set. The user selects a focus, i.e., object of interest, from a set of pre-defined features. Our system automatically determines the most expressive view on this feature. A characteristic viewpoint is estimated by a novel information-theoretic framework which is based on the mutual information measure. Viewpoints change smoothly by switching the focus from one feature to another one. This mechanism is controlled by changes in the importance distribution among features in the volume. The highest importance is assigned to the feature in focus. Apart from viewpoint selection, the focusing mechanism also steers visual emphasis by assigning a visually more prominent representation. To allow a clear view on features that are normally occluded by other parts of the volume, the focusing for example incorporates cut-away views.
本文介绍了一种用于自动聚焦体积数据集中特征的概念。用户从一组预定义特征中选择一个焦点，即感兴趣的对象。我们的系统会自动确定该特征最具表现力的视图。通过一种基于互信息度量的新颖信息论框架来估计特征视角。通过将焦点从一个特征切换到另一个特征，视角会平滑变化。这种机制由体积中特征重要性分布的变化控制。最高重要性被赋予焦点处的特征。除了视角选择外，聚焦机制还通过分配视觉上更突出的表示来引导视觉强调。为了能清晰查看通常被体积其他部分遮挡的特征，例如聚焦时会采用剖视图。
