Content-Based Multimedia Data Retrieval on Cluster System Environment

In this paper, we propose a data partitioning scheme for contentbased retrieval from a multimedia database using a heterogeneous cluster system. The proposed parallel unified model [6] was used to represent multimedia data. All types of multimedia data in the unified model are represented by k-dimensional(k-d) signals. Each dimension of k-d data is separated into small blocks and then formed into a hierarchical multidimensional tree structure, called a k-tree. The parallel version of k-tree model was introduced in [7]. The previous experimental results show the huge reduction of retrieval time on a cluster of homogeneous workstations. In this paper, we extend our parallel model to a heterogeneous cluster system environment. We demonstrate the experimental results of using a parallel retrieval algorithm for the k-tree unified model on a cluster of heterogeneous system connected via a network. We use system characteristics to help in partitioning the data and balancing the loads of the processors. The experiments of the model with load balancing shows a significant reduction of retrieval time while maintaining the quality of perceptual results.