An exact match retrieval scheme based upon principal component analysis

Abstract

How to perceive spatial relationships among objects in a picture is an important criterion to retrieve the pictures in a symbolic picture database. In 1991, Chang proposed an indexing structure, called the 9DLT matrix, to logically represent symbolic pictures by pairwise spatial relationships. The corresponding 9DLT matrix preserves the spatial relationships among objects in a symbolic picture. In this paper, a match retrieval of pictures in a symbolic picture database by the use of 9DLT matrix is proposed. The proposed scheme is based upon a famous multivariant statistic technique — principal component analysis. By the proposed scheme, it requires O(log n) search time in the worst case for the exact match retrieval of symbolic pictures, where n is the total number of pictures in the data base.