A Temporal and Multi-Resolution Visualization System

[[abstract]]在這篇論文中，我們提出一個基於力導向演算法的多解析度的視覺化方式，並用圖形與階層化來表示動態的資料集合。考慮到人類視覺感官系統的侷限，抽取出最重要的資料、或是使用者設定的資料來表現出資訊是很合理的。為了能提升使用者辨識效率，限制物件的數量也是非常重要的。為了更平滑地表示動態graph的動畫效果，我們採用了類似果凍狀的元球的技術來模擬graph。在我們的系統中，將上千筆節點(資料)以及連結(關係)透過群組化與圖像化來顯示，好讓使用者更容易地辨認。我們的系統除了可以透過互動式2D/3D圖形介面並動態地顯示資訊的多階層結構以及資訊的趨勢，同時也可以顯示資料的細節以及摘要。更甚者，應用3D立體的技巧，如透視、陰影、聚焦的方法，可以提升使用者理解資料的效率。 在實驗中，我們展示了使用了我們的系統來瀏覽Google Plus所提供的網路言論。使用者可以透過隨著時間推移的動態圖以及互動式介面，來輕易的理解被公開的言論趨勢以及之間的關係。[[abstract]]In this paper, we propose a multi-resolution visualization approach based on a force-directed algorithm to graphically and hierarchically represent dynamic data sets. Considering the limitations of the human visual cognitive system, it is reasonable to extract the most important or user-specified information for representation. Furthermore, in order to improve the efficiency of human recognition, it is important to restrict the number of objects in a graph. To smoothly animate changes of a dynamic graph, we adopt a jelly-like metaball technique. Our system clusters and graphically represents thousands of nodes (data) and links (relationships) in a novel way that allows a user to recognize the information visually. Our system displays information in a hierarchical format and evolving trends are dynamically visualized through an interactive 2D/3D graphical interface while simultaneously summarizing and displaying data detail. Furthermore, user perception can be enhanced by adopting human 3D stereopsis skills such as perspective, chiaroscuro, and focusing methods. Our system is demonstrated by exploring internet comments from Google Plus. Users can easily utilize the time-varying graph of public comments trends and their interrelationships over time through the interface.[[note]]博