Query refinement for Internet multimedia information retrieval using keywords and low-level features

The main obstacle in realizing semantic-based image retrieval is front the web that semantic description of an image is difficult to capture in low-level features. Text based keywords can be generated from web documents to capture semantic information for narrowing down the search space. We use an effective approach to integrate keywords, which is extracted from Web documents and low-level features such as color-texture features to take advantage of their complementing strengths. Experimental results show that the keywords can be replaced with low-level features for successful query refinement and improved precision of retrieval.