Semantic Query Expansion Using Knowledge Based for Images Search and Retrieval

The falling prices of multimedia and storage devices make almost everyone to act like a professional to capture photo and archive them for later use. Without efficient retrieval methods the search of images in large collections can become a painstaking work. Most of the traditional image search engines rely on keyword-based annotations which lacks the query semantic space equivalent to the annotation semantic space, because of the difficulty in describing the same concepts with other keywords. In this paper, we propose a novel approach for the query expansion using lexical and commonsensical knowledge based like WordNet and ConceptNet, which will not only fill the gap in the semantic space between user query and annotation but will also provide an opportunity to discard the less important words from the query semantic space. For evaluation we have selected LabelMe data sets, which is openly available for researcher.