Pragmatic Security for Efficient Web Search Using Amalgamation

Natural language processing alone does not improve the accuracy of information retrieval systems. If the focus is shifted to short phrases rather than full documents, the situation becomes extremely different. The proposed system uses the combination of game theory, natural language processing technique and steganography to obtain high accuracy retrieval of images which have been annotated with a descriptive textual caption. The natural language techniques also allow additional contextual information to be derived from the relation between the query and the caption along with high security to store the information in a database, which can help users to understand the overall protected collection of retrieval results.