Adaptive Two-Phase Tree-Based Spatial Data Structures and Their Application to Image Compression

Abstract

Given a binary image, this paper first observes that the conventional tree-based spatial data structures (SDSs) spend some amount of memory for storing the external nodes and internal nodes near leaves. Next, we present an adaptive two-phase (ATP) representation to reduce the memory requirement. In addition, some geometric operations, such as computing the area and centroid, on the proposed ATP representation are also discussed. Experimental results show that not only the proposed ATP representation can reduce the memory requirement when compared to the existing ones, but also can lead to faster concerning geometrical operations. Finally, the proposed results are extended to compress gray images.