Ad hoc image retrieval using hierarchical semantic-based index

Content-based image retrieval is a challenging problem in mobile ad hoc networks due to the multiple limitations such as network bandwidth, infrastructure-free nature, and node mobility. The traditional systems employ either centralized or flooding strategies, which may result in low fault tolerance or high search cost. In this paper, we propose a decentralized non-flooding retrieval scheme in multi-hop ad hoc networks - semantic ad hoc image retrieval (SAIR). The novelty of SAIR stems from several factors including: (1) representation of image contents using first-order logic expressions; (2) clustering mobile nodes based on their data contents; and (3) performing content-based image retrieval within a reduced scope of mobile nodes. Through extensive simulations, we show that relative to the flooding strategy, SAIR can retrieve the semantically most similar image objects by accessing only a small portion of the mobile nodes with lower search cost. Moreover, it is scalable to large network sizes and large number of data objects.