Neural Based Approach for Fast Retrieval of Symbolic Images in Database

The rapid growth of image databases results in new requirement for fast image indexing and retrieval. In this paper we use Correlation Matrix Memory (CMM) as data structure for fast image storage and retrieval. We use Triangular Spatial Relation (TSR) to represent spatial relations among image iconic objects. The proposed model is faster than other existing models in both partial and exact match retrieval (¿(1)).