In this paper, we describe a distributed paged-object server to efficiently support the storage management for new generation database management systems. This storage server is based on distributed client/server architecture, and allows clients to directly map database files onto distributed shared virtual memory (DSVM). In this architecture, there is a server at each site and the server only supports the clients in the same site. This improves performance utilizing client machine resources and offloading the shared resourcer server machine and the network. Therefore, this architecture may avoid drawbacks such as bottleneck in a traditional centralized client/server architecture, and also may reduces network traffic and improve performance efficiency of remote file access using Net File System (NFS). Moreover, this architecture allows distributed shared objects to reside and execute anywhere and to be used by any clients on the network. A prototype system (called WAKASHI/D) is implemented under the Mach operating system. The distributed, shared, and transactional virtual memory that the system supports is either volatile or persistent and they can be accessed by user applications in a uniform way. This paper also presents a performance evaluation and analysis of WAKASHI/D to compare its centralized version WAKASHI/C and demonstrates that the distributed server has substantial performance benefits.

This paper proposes a technique for indexing, clustering and retrieving images based on their edge features. In this technique, images are decomposed into several frequency bands using the Haar wavelet transform. From the one-level decomposition sub-bands an edge image is formed. Next, the higher order auto-correlation function is applied on the edge image to extract the edge features. These higher order autocorrelation features are normalized to generate a compact feature vector, which is invariant to shift, image size. We used direction cosine as measure of distance not to be influenced by difference of each image's luminance. Then, these feature vectors are clustered by a self-organizing map (SOM) based on their edge feature similarity. The performed experiments show higher precision and recall of this technique than traditional ways in clustering and retrieving images in a large image database environment.

In this paper, we present a novel approach for efficient search of high-dimensional databases, such as video shots. The idea is to map feature vectors from the high-dimensional feature space into a point in a low-dimensional distance space. Then, a spatial access method, such as an R-tree, is used to cluster these points based on their distances in the low-dimensional space. Our mapping method, called topological mapping, guarantees no false dismissals in the result of a query. However, the result of a query might contain some false alarms. Hence, two refinement steps are performed to remove these false alarms. Comparative experiments on a database of video shots show the superior efficiency of the topological mapping method over other known methods.

When a set of objects is shared among several applications, multiple implementations for the set are required in order to suit each application as much as possible. Furthermore, if a set of objects could have multiple implementations, the following issues arise: (1) how to select the best implementation when processing queries on the set, and (2) how to propagate updates on an implementation of the set to the others. In this paper we propose a mechanism of multiple implementations for a set, and also give a solution for the latter issue. In the proposal a set can be of multiple types, and each of the types corresponds to an implementation already contained within the set. Update propagation can be achieved by a rewriting technique at compilation time. We also present a performance study in which the feasibility and effectiveness of our proposal were examined.

This paper gives an overview of research activities on high performance databases in Japan. It focuses on parallel algorithms for relational databases and data mining, parallel approaches for object-oriented databases, and parallel disk systems. Studies surveyed in this paper are carried out mainly by database researchers in Japanese universities under the Grant-in-Aid for Scientific Research (1996-1998).

In light of the increasing number of computer applications that rely heavily on multimedia data, the database community has focused on the management and retrieval of multidimensional data. Nearest Neighbor queries (NN queries) have been widely used to perform content-based retrieval (e.g., similarity search) in multimedia applications. Incremental NN (INN) query is a kind of NN queries and can also be used when the number of the NN objects to be retrieved is not known in advance. This paper points out the weaknesses of the existing INN search algorithms and proposes a new one, called Batch-Incremental Nearest Neighbor search algorithm (denoted B-INN search algorithm), which can be used to process the INN query efficiently. The B-INN search algorithm is different from the existing INN search algorithms in that it does not employ the priority queue that is used in the existing INN search algorithms and is very CPU and memory intensive for large databases in high-dimensional spaces. And it incrementally reports b(b > 1) objects simultaneously (Batch-Incremental), whereas the existing INN search algorithms report the neighbors one by one. In order to implement the B-INN search, a new search (called k-d-NN search) with a new pruning strategy is proposed. Performance tests indicate that the B-INN search algorithm clearly outperforms the existing INN search algorithms in high-dimensional spaces.

Recently, two approaches investigated indexing and retrieving videos. One approach utilized the visual features of individual objects, and the other approach exploited the spatio-temporal relationships between multiple objects. In this paper, we integrate both approaches into a new video model, called the Visual-Spatio-Temporal (VST) model to represent videos. The visual features are modeled in a topological approach and integrated with the spatio-temporal relationships. As a result, we defined rich sets of VST relationships which support and simplify the formulation of more semantical queries. An intuitive query interface which allows users to describe VST features of video objects by sketch and feature specification is presented. The conducted experiments prove the effectiveness of modeling and querying videos by the visual features of individual objects and the VST relationships between multiple objects.