Indexing binary images using quad-tree decomposition

The tree structure is introduced to specify block-oriented decomposition of database images. These decomposition structures offer a fundamental data model for specifying image content in large image databases. We propose in this paper a new indexing and classification method based on the use of the quad-tree structure. 3D objects are represented by their silhouettes and codified following the filling rate of each quadrant at different levels of the quad-tree subdivision. It is shown that three decomposition levels are sufficient to efficiently index all the images of the database. We propose also a modified linear codification for silhouettes. Our approach allows the reduction of the processing time and the memory space to store images codification following the structure of the quad-tree.