Search

We describe a re-ranking method called dual rank to improve web image retrieval by clustering and reordering the images retrieved from an image search engine. General image retrieval exploits text and links the structure or little visual information to ‘understand’ the content of the web images, and usually lack the discriminative power to deliver visually diverse search results. The framework of dual rank is composed of inter-cluster rank and intra-cluster rank. To address the clustering problem, we first utilize a multipartite graph model to represent images and features, then formulate clustering as a constrained multi-objective optimization problem, which can be efficiently solved by semi-definite programming. We employ inter-cluster ranking function and intra-cluster ranking function to rank clusters and images, respectively. A representative image is selected from each cluster which together forms a diverse result set and as the optimal results for a query. We fuse different image features (text, colour, shape, texture, etc.) to improve the effect of clustering. Our method is evaluated against a standard search engine and significant improvements are reported in terms of Mean average precision, D@n and user experience. [ABSTRACT FROM AUTHOR]

In this paper, we present a hybrid relevant-diverse image re-ranking approach that combines the strengths of two previous methods: the reciprocal election algorithm proposed by R. van Leuken et al. and the greedy search algorithm proposed by T. Deselaers et al. Our approach is a cluster-based re-ranking method. We select several candidate representatives based on the reciprocal election algorithm and employ a bounded greedy search algorithm to find the most relevant-diverse one as the cluster representative. We fuse multiple features to calculate image similarity, including colour and shape and especially topic content features, and discuss the benefits of integrating different features. In addition, we quantitatively evaluate our approach on a real-world Web image data set and obtain results which outperform the state of the art. [ABSTRACT FROM AUTHOR]

Managing transactions is very important in distributed databases in order to preserve data consistency and reliability of the systems. This paper presents a new design to manage transactions on neighbour replication in a distributed database system. We address how to build a reliable system for managing transactions on a neighbour replication grid (NRG) in terms to preserve the data consistency and support fault-tolerance. We first recall the model and technique of NRG that impose neighbours binary vote assignment to its logical grid structure on data copies. We extend our work in managing transactions to normal and failure cases. Finally, the implementation of the system is presented. [ABSTRACT FROM AUTHOR]