A Generalized Non-redundant 2D String Representation for Symbolic Pictures

Abstract

In this paper, we generalize the non-redundant(NR)2D string representation for symbolic pictures to reduce the storage needed and to avoid ambiguity. In image database systems, we can use 2D string representation to efficiently construct iconic indexes for pictures. But,2D string requires storage for redundant elements. This shortcoming is overcome by the NR2D string representation. However, the ambiguous problem is arised under the NR2D string representation. The NR2D string representation confuses the switch components. Besides,the retrieval efficiency of image queries is degraded. In our generalized NR2D string representation, the different switch components are impossible to be transformed to the same 2D string representation. And thus our method can improve the retrieval efficiency of image queries. So the ambiguous problem can be overcome.