以二維C-字串(2D C-string)為基礎之圖形資料結構已被用來表示符號圖(Symbolic picture)。它提供了以非常自然的方式來建立圖示索引(Iconic indexes),而且物件間之空間關係能根據這些字串利用空間推理(Spatial reasoning)的方式而獲得。但這些字串產生之方法皆是根據各物件之最小邊界矩形(Minimum bounding rectangles,MBR)而產生,因此,推論出來之空間關係與真實物件之空間關係存有相當差異。在本論文中,我們將提出一個減少此差異之方法,將較大之物件分解成較小物件再取其MBR而產生二維C-字串,而據此推論出差異較小,較接近真實物件之空間關係。產生此種二維C-字串以及其空間推理之方法將被提出。 The pictorial data structures based on 2D C-strings have been used to represent symbolic pictures. It provides a natural way to construct the iconic indexes for pictures. The spatial relationships between any two objects can be obtained by using spatial reasoning based on these strings. However, these strings are produced according to the minimum bounding rectangles (MBR) of objects. Hence, there is some difference between the inferred spatial relationships and that between the real objects. In this paper, we propose an approach that can reduce the difference. Bigger objects are decomposed into smaller objects first, and then MBRs are generated for the smaller objects. Finally, 2D C-strings are generated and the spatial relationships with less difference can be obtained by spatial reasoning. The methods of producing the 2D C- strings based on decomposed objects and the spatial reasoning according to these strings are proposed.
中華民國八十四年全國計算機會議論文集第二冊=Proceedings of National Computer Symposium 1995 Republic of China Volume 2，頁714-720