BoVW Model in Content Based Image Retrieving

A scalable approach for sharing multimedia data across the network is the Peer-to-peer networking. For a large amount of data which is distributed among different nodes, to perform content-based retrieval in peer-to-peer networks is an important but challenging problem. But many of the existing methods are focus on the indexing high dimensional visual features. This type of methods has scalability limitations. By using the bag-of-visual words model found a scalable approach for content-based image retrieval in peer-to-peer networks. By making a comparison with centralized ones, the key problem is to efficiently and precisely obtain a codebook which is global. For a peer-to-peer network, which is often evolves dynamically, and that makes a static codebook which is less effective for the retrieval tasks. Therefore, have to propose a dynamic codebook updating method. This is by optimization of the mutual information between the codebook which is produced and important information. To manage the different code words the workload balanc914e among nodes is necessary. Indexing pruning techniques are used for the improvement in the retrieval process and for reducing the network cost.