提升三維多點互動系統使用者辨識能力之多圖騰演算法開發

Abstract

近年來3D顯示科技的蓬勃發展，使得影像跳脫傳統框架，栩栩如生地出現在我們眼前。此外2D觸控面板的爆炸性成長，使得多點觸控改變了使用者對於人機介面的使用習慣與期待。鑒於立體影像已蔚為風潮，加上智慧型手機與平板電腦的普及，未來將兩者結合之3D智慧型手機與3D平板電腦已然成為趨勢。 然而要使3D技術普及化，除了3D顯示設備外，擁有一個友善的人機互動介面更是推動此項技術的關鍵，因此3D互動系統將成為下世代的關鍵技術。為了達成平面與三維空間的連續互動介面，我們選擇了內嵌光感測元件的架構，提出適用於薄型可攜式電子顯示器之多點、以及多人三維立體近距離互動系統。 在以往的3D多點/多人互動系統中，使用者間部分重疊的情況始終為一大困擾，因此，經由所提出的演算法，包括時序性互動演算法、多圖騰式演算法等，多位使用者可同時間的與系統進行三維互動並被成功的分辨。最後，所提出的演算法被證實於內嵌光感測元件的4吋面板上。多環形圖騰演算法提升了使用者的辨識能力，解決了使用者間40%之重疊情形，同時獲得個別使用者之二維的位置(x,y)以及深度資訊(z)；另外，藉由多T形圖騰演算法，不僅是使用者的三維資訊(x,y,z)，其傾斜與旋轉角度(θ,φ)等資訊也能夠被準確地截取出來。

Recently, the widespread of 2D multi-touch user interfaces has become a part of our lives. Meanwhile, the increasing popularity in 3D images dramatically changes the user experience. Therefore, by the fact of great mass fervor in smart phone and tablet computer, 3D smart phone and 3D tablet is undoubted the battle field of next generation technology. However, a portable 3D device cannot be all-pervading if there is no friendly 3D multi-interactive user interfaces. Hence, in order to establish a continuous interaction space from 2D surface to 3D near field distance, the embedded optical sensor based structure is chosen in our approaches. However, for multi-user/multi-touch interactive systems, the users’ overlapping has been a serious issue in user recognition. Therefore, by the proposed methods, the sequential-lighting method and the multi-mark method, user overlapping issue might be solved. Three algorithms, the sequential-lighting algorithm, multi-ring mark algorithm, and the multi-T mark algorithm, and the accordant light marks are proposed to provide distinguishable 3D user interfaces. Finally, the feasibility of the proposed algorithms was demonstrated on a 4” panel with in-cell optical sensors. The ability of user identification can be improved to 40% overlapping by the multi-ring mark algorithm. Meanwhile, 5-axis information (x, y, z, θ, and φ) including 3D coordinate and orientation can be obtained through the multi-T mark algorithm.